diff --git a/backend/agenteval/evaluation/report_render.py b/backend/agenteval/evaluation/report_render.py
index 12ebac9..88a5762 100644
--- a/backend/agenteval/evaluation/report_render.py
+++ b/backend/agenteval/evaluation/report_render.py
@@ -7,6 +7,7 @@ show) belong to generation; this module only formats.
 """
 
 import json
+from datetime import datetime
 from typing import Any, Optional
 
 from jinja2 import Template
@@ -213,25 +214,164 @@ def _render_analysis_lines(analysis: dict[str, Any], scenario_names: dict[str, s
     return lines
 
 
+_CAMPAIGN_STATUS_LABELS = {
+    "planned": "计划中", "running": "进行中", "completed": "已完成",
+    "cancelled": "已取消", "failed": "失败",
+}
+_TREND_LABELS = {"improving": "改善", "stable": "平稳", "regressing": "退化"}
+_EVOLUTION_LABELS = {"new": "新增", "persisting": "持续", "resolved": "消解"}
+_TRACKING_LABELS = {"addressed": "已落实", "partial": "部分落实", "unaddressed": "未落实", "new": "新增"}
+
+
+def _offset(seconds: float) -> str:
+    """Human-readable window offset: 86400 → 24h, 1800 → 30m, 45 → 45s."""
+    s = int(round(seconds))
+    if s % 3600 == 0:
+        return f"{s // 3600}h"
+    if s % 60 == 0:
+        return f"{s // 60}m"
+    return f"{s}s"
+
+
+def _dt(iso: Optional[str]) -> str:
+    """ISO timestamp → `2026-08-01 08:00`; unparseable input passes through."""
+    if not iso:
+        return "-"
+    try:
+        return datetime.fromisoformat(iso.replace("Z", "+00:00")).strftime("%Y-%m-%d %H:%M")
+    except ValueError:
+        return iso
+
+
+def _diff_cell(metric: str, pair: dict[str, Any]) -> str:
+    """One metric cell of the diff table: `基线 → 本期（±delta）`."""
+
+    def _val(v: Optional[float]) -> str:
+        if v is None:
+            return "—"
+        return _ms(v) if metric == "avg_latency_ms" else _pct(v)
+
+    delta = pair.get("delta")
+    if delta is None:
+        delta_text = "—"
+    elif metric == "avg_latency_ms":
+        delta_text = f"{delta:+.1f}ms"
+    else:
+        delta_text = f"{delta * 100:+.1f}pp"
+    return f"{_val(pair.get('baseline'))} → {_val(pair.get('current'))}（{delta_text}）"
+
+
+def _render_comparison_lines(comparison: dict[str, Any], scenario_names: dict[str, str]) -> list[str]:
+    """Appendix for the campaign export: narrative + mechanical metric diff.
+
+    ``comparison`` carries the completed narrative result plus context resolved
+    on the read path (baseline name/times, model name, generated-at, diff).
+    """
+
+    def _name(sid: str) -> str:
+        return scenario_names.get(sid, sid[:8])
+
+    result = comparison.get("result") or {}
+    lines = ["", "## 周期对比", ""]
+    meta: list[str] = []
+    if comparison.get("baseline_name"):
+        base = f"基线：「{comparison['baseline_name']}」"
+        if comparison.get("baseline_completed_at"):
+            base += f"（完成于 {_dt(comparison['baseline_completed_at'])}）"
+        meta.append(base)
+    if comparison.get("model_name"):
+        meta.append(f"分析模型：{comparison['model_name']}")
+    if comparison.get("updated_at"):
+        meta.append(f"生成于：{_dt(comparison['updated_at'])}")
+    if meta:
+        lines += [" · ".join(meta), ""]
+
+    trend = _TREND_LABELS.get(result.get("trend"), "平稳")
+    lines += [f"**趋势**：{trend} — {result.get('summary') or '—'}", ""]
+
+    diff = comparison.get("metric_diff")
+    if diff:
+        lines += [
+            "### 指标变化",
+            "",
+            "| 维度 | 通过率 | 可用性 | 平均时延 |",
+            "|------|------|------|------|",
+        ]
+
+        def _row(label: str, block: dict[str, Any]) -> str:
+            return (
+                f"| {label} | {_diff_cell('pass_rate', block['pass_rate'])} | "
+                f"{_diff_cell('availability', block['availability'])} | "
+                f"{_diff_cell('avg_latency_ms', block['avg_latency_ms'])} |"
+            )
+
+        lines.append(_row("整窗（总体）", diff["overall"]))
+        for s in diff.get("scenarios") or []:
+            label = s.get("scenario_name") or _name(s.get("scenario_id", ""))
+            lines.append(_row(label, s))
+        lines.append("")
+
+    evolution = result.get("problem_evolution") or []
+    if evolution:
+        lines += ["### 问题演变", ""]
+        for p in evolution:
+            status = _EVOLUTION_LABELS.get(p.get("status"), "持续")
+            names = "、".join(_name(sid) for sid in p.get("scenario_ids") or [])
+            head = f"- **[{status}] {p.get('title', '')}**"
+            if names:
+                head += f"（场景：{names}）"
+            lines.append(head)
+            if p.get("detail"):
+                lines.append(f"  {p['detail']}")
+        lines.append("")
+
+    tracking = result.get("suggestion_tracking") or []
+    if tracking:
+        lines += ["### 建议落实情况", ""]
+        for t in tracking:
+            status = _TRACKING_LABELS.get(t.get("status"), "未落实")
+            lines.append(f"- **[{status}] {t.get('text', '')}**")
+            if t.get("note"):
+                lines.append(f"  {t['note']}")
+        lines.append("")
+    return lines
+
+
+def _window_line(report: dict[str, Any]) -> str:
+    """Human-readable window line with the 正式线 / 加速调试线 wording."""
+    window = report.get("window_seconds") or 0
+    scale = float(report.get("time_scale") or 1)
+    if scale == 1:
+        return f"**窗口**: {_offset(window)}（正式线）"
+    wall = f"，压缩后实际耗时约 {_offset(window / scale)}"
+    return f"**窗口**: {_offset(window)}（加速调试线 ×{scale:g}{wall}）"
+
+
 def render_campaign_markdown(
     report: dict[str, Any],
     *,
     analysis: Optional[dict[str, Any]] = None,
+    comparison: Optional[dict[str, Any]] = None,
+    target_name: Optional[str] = None,
     scenario_names: Optional[dict[str, str]] = None,
 ) -> str:
     """Render a dual-axis campaign report dict as Markdown.
 
-    ``analysis`` is the stored 智能分析 result dict (completed only); when
-    absent the export is identical to the pre-analysis format.
+    ``analysis`` and ``comparison`` are the stored 智能分析 / 周期对比 results
+    (completed only); when absent the corresponding appendix is omitted
+    entirely.
     """
     s = report["summary"]
+    status = _CAMPAIGN_STATUS_LABELS.get(report.get("status"), report.get("status") or "-")
+    target = target_name or report.get("target_id") or "-"
     lines: list[str] = [
         f"# 活动周期报告 — {report['name']}",
         "",
-        f"**状态**: {report['status']}  ",
-        f"**窗口**: {report['window_seconds']}s（倍速 {report['time_scale']}）  ",
-        f"**开始时间**: {report['started_at'] or '-'}  ",
-        f"**完成时间**: {report['completed_at'] or '-'}  ",
+        f"**评测对象**: {target}  ",
+        f"**状态**: {status}  ",
+        _window_line(report) + "  ",
+        f"**开始时间**: {_dt(report.get('started_at'))}  ",
+        f"**完成时间**: {_dt(report.get('completed_at'))}  ",
         "",
         "## 汇总",
         "",
@@ -245,12 +385,12 @@ def render_campaign_markdown(
         "",
         "## 时间趋势",
         "",
-        "| 时段(秒) | 运行数 | 通过率 | 可用性 | 时延 |",
+        "| 时段 | 运行数 | 通过率 | 可用性 | 时延 |",
         "|------|------|------|------|------|",
     ]
     for b in report["time_trend"]:
         lines.append(
-            f"| {b['start_seconds']:.0f}–{b['end_seconds']:.0f} | {b['run_count']} | "
+            f"| {_offset(b['start_seconds'])}–{_offset(b['end_seconds'])} | {b['run_count']} | "
             f"{_pct(b['pass_rate'])} | {_pct(b['availability'])} | {_ms(b['avg_latency_ms'])} |"
         )
     lines += [
@@ -267,4 +407,6 @@ def render_campaign_markdown(
         )
     if analysis:
         lines += _render_analysis_lines(analysis, scenario_names or {})
+    if comparison:
+        lines += _render_comparison_lines(comparison, scenario_names or {})
     return "\n".join(lines)
diff --git a/backend/agenteval/web/routers/campaigns.py b/backend/agenteval/web/routers/campaigns.py
index 4eb2fb9..831b73e 100644
--- a/backend/agenteval/web/routers/campaigns.py
+++ b/backend/agenteval/web/routers/campaigns.py
@@ -134,9 +134,39 @@ async def get_campaign_report_markdown(campaign_id: str, session: Session = Depe
     scenario_names = {s.id: s.name for s in ScenarioRepository(session).list_all()}
     analysis_row = CampaignAnalysisRepository(session).get_by_campaign(campaign_id)
     analysis = analysis_row.get_result() if analysis_row and analysis_row.status == "completed" else None
+    target = TargetRepository(session).get(campaign.target_id)
+    target_name = target.name if target else None
+
+    comparison = None
+    cmp_row = CampaignPeriodComparisonRepository(session).get_by_campaign(campaign_id)
+    if cmp_row is not None and cmp_row.status == "completed" and cmp_row.get_result():
+        baseline = CampaignRepository(session).get(cmp_row.baseline_campaign_id)
+        metric_diff = (
+            compute_metric_diff(
+                build_campaign_report_dict(baseline, session),
+                build_campaign_report_dict(campaign, session),
+            )
+            if baseline is not None
+            else None
+        )
+        model_cfg = ModelConfigRepository(session).get(cmp_row.model_config_id) if cmp_row.model_config_id else None
+        model_label = f"{model_cfg.name}（{model_cfg.model_name}）" if model_cfg and model_cfg.model_name else (
+            model_cfg.name if model_cfg else None
+        )
+        comparison = {
+            "result": cmp_row.get_result(),
+            "baseline_name": baseline.name if baseline else None,
+            "baseline_completed_at": iso_utc(baseline.completed_at) if baseline else None,
+            "model_name": model_label,
+            "updated_at": iso_utc(cmp_row.updated_at),
+            "metric_diff": metric_diff,
+        }
+
     md = render_campaign_markdown(
         generate_campaign_report(campaign, runs, scenario_names=scenario_names),
         analysis=analysis,
+        comparison=comparison,
+        target_name=target_name,
         scenario_names=scenario_names,
     )
     return Response(
diff --git a/tests/integration/test_campaign_comparison_api.py b/tests/integration/test_campaign_comparison_api.py
index 6d9ac64..5ddd8bd 100644
--- a/tests/integration/test_campaign_comparison_api.py
+++ b/tests/integration/test_campaign_comparison_api.py
@@ -407,3 +407,60 @@ async def test_post_comparison_rerun_upserts_without_new_row(client, seeded_db,
         select(CampaignPeriodComparisonDB).where(CampaignPeriodComparisonDB.campaign_id == current_id)
     ).all()
     assert len(rows) == 1
+
+
+# ── Markdown 导出纳入周期对比 ──────────────────────────────────────────────
+
+
+async def test_markdown_export_includes_completed_comparison(client, seeded_db):
+    _seed_analysis_default(seeded_db)
+    baseline_id, current_id = await _two_period_setup(client, seeded_db)
+    _seed_run(seeded_db, "run-base", baseline_id, pass_rate=0.5, latency=800.0)
+    _seed_run(seeded_db, "run-cur", current_id, pass_rate=0.9, latency=500.0)
+    _complete_comparison_row(seeded_db, current_id, baseline_id)
+
+    resp = await client.get(f"/api/campaigns/{current_id}/report/markdown")
+    assert resp.status_code == 200
+    md = resp.text
+
+    # 排版重优化：头部评测对象名、中文化状态、正式线窗口
+    assert "**评测对象**: mock-target" in md
+    assert "**状态**: 已完成" in md
+    assert "**窗口**: 24h（正式线）" in md
+
+    # 周期对比附录：元信息 + 趋势 + 机械 diff 表
+    assert "## 周期对比" in md
+    assert "基线：「上期」" in md
+    assert "分析模型：analysis-cfg（m）" in md
+    assert "**趋势**：改善 — 通过率提升" in md
+    assert "| 整窗（总体） | 50.0% → 90.0%（+40.0pp） | 100.0% → 100.0%（+0.0pp） | 800ms → 500ms（-300.0ms） |" in md
+    assert "**[消解] 答非所问**（场景：mock-scenario）" in md
+    assert "**[已落实] 保持**" in md
+
+    # 智能分析附录在前，周期对比紧随其后
+    assert md.index("## 智能分析") < md.index("## 周期对比")
+
+
+async def test_markdown_export_without_comparison_row_has_no_section(client, seeded_db):
+    _seed_analysis_default(seeded_db)
+    _, current_id = await _two_period_setup(client, seeded_db)
+    _seed_run(seeded_db, "run-cur", current_id, pass_rate=0.9, latency=500.0)
+
+    resp = await client.get(f"/api/campaigns/{current_id}/report/markdown")
+    assert resp.status_code == 200
+    assert "周期对比" not in resp.text
+
+
+async def test_markdown_export_skips_non_completed_comparison(client, seeded_db):
+    _seed_analysis_default(seeded_db)
+    baseline_id, current_id = await _two_period_setup(client, seeded_db)
+    seeded_db.add(CampaignPeriodComparisonDB(
+        campaign_id=current_id, baseline_campaign_id=baseline_id,
+        status="failed", model_config_id="mc-1", triggered_by="manual",
+        error="boom",
+    ))
+    seeded_db.commit()
+
+    resp = await client.get(f"/api/campaigns/{current_id}/report/markdown")
+    assert resp.status_code == 200
+    assert "周期对比" not in resp.text
diff --git a/tests/unit/test_report_render.py b/tests/unit/test_report_render.py
index 19f245c..9619cc5 100644
--- a/tests/unit/test_report_render.py
+++ b/tests/unit/test_report_render.py
@@ -184,7 +184,7 @@ def test_render_campaign_markdown_summary_and_axes():
 def test_render_campaign_markdown_empty_bucket_dashes():
     md = render_campaign_markdown(_campaign_report())
     # bucket 1 has no runs: pass_rate/availability/latency all render as —
-    assert "| 3600–7200 | 0 | — | — | — |" in md
+    assert "| 1h–2h | 0 | — | — | — |" in md
 
 
 def _analysis() -> dict:
@@ -246,3 +246,103 @@ def test_render_campaign_markdown_analysis_falls_back_to_id_prefix():
 def test_render_campaign_markdown_without_analysis_unchanged():
     md = render_campaign_markdown(_campaign_report())
     assert "智能分析" not in md
+
+
+def test_render_campaign_markdown_header_readability():
+    md = render_campaign_markdown(_campaign_report(), target_name="客服机器人")
+    assert "**评测对象**: 客服机器人" in md
+    assert "**状态**: 已完成" in md
+    assert "**窗口**: 2h（正式线）" in md
+    assert "**开始时间**: 2026-07-30 00:00" in md
+    assert "**完成时间**: 2026-07-30 02:00" in md
+    assert "| 0h–1h | 2 | 75.0% | 100.0% | 150ms |" in md
+
+
+def test_render_campaign_markdown_header_falls_back_to_target_id():
+    md = render_campaign_markdown(_campaign_report())
+    assert "**评测对象**: t-1" in md
+
+
+def test_render_campaign_markdown_accelerated_line():
+    report = _campaign_report()
+    report["window_seconds"] = 86400
+    report["time_scale"] = 4.0
+    md = render_campaign_markdown(report)
+    assert "**窗口**: 24h（加速调试线 ×4，压缩后实际耗时约 6h）" in md
+
+
+def _comparison() -> dict:
+    """A hand-built dict matching the comparison context the export endpoint resolves."""
+    return {
+        "result": {
+            "trend": "regressing",
+            "summary": "整体质量下滑，售后场景恶化",
+            "problem_evolution": [
+                {
+                    "status": "persisting",
+                    "title": "售后答非所问",
+                    "scenario_ids": ["s-1"],
+                    "detail": "问题仍未收敛",
+                },
+                {"status": "resolved", "title": "响应偏慢", "scenario_ids": [], "detail": ""},
+            ],
+            "suggestion_tracking": [
+                {"status": "partial", "text": "补充售后知识库", "note": "仅覆盖部分问题"},
+                {"status": "new", "text": "新增建议：监控时延", "note": ""},
+            ],
+        },
+        "baseline_name": "上一期巡检",
+        "baseline_completed_at": "2026-07-29T02:00:00+00:00",
+        "model_name": "qwen-max",
+        "updated_at": "2026-07-30T03:00:00+00:00",
+        "metric_diff": {
+            "overall": {
+                "pass_rate": {"baseline": 0.8, "current": 0.75, "delta": -0.05},
+                "availability": {"baseline": 1.0, "current": 0.5, "delta": -0.5},
+                "avg_latency_ms": {"baseline": 120.0, "current": 150.0, "delta": 30.0},
+            },
+            "scenarios": [
+                {
+                    "scenario_id": "s-1",
+                    "scenario_name": "售后场景",
+                    "pass_rate": {"baseline": 0.7, "current": 0.6, "delta": -0.1},
+                    "availability": {"baseline": None, "current": None, "delta": None},
+                    "avg_latency_ms": {"baseline": 100.0, "current": 140.0, "delta": 40.0},
+                },
+            ],
+        },
+    }
+
+
+def test_render_campaign_markdown_appends_comparison_section():
+    md = render_campaign_markdown(
+        _campaign_report(),
+        comparison=_comparison(),
+        scenario_names={"s-1": "售后场景"},
+    )
+    assert "## 周期对比" in md
+    assert "基线：「上一期巡检」（完成于 2026-07-29 02:00）" in md
+    assert "分析模型：qwen-max" in md
+    assert "生成于：2026-07-30 03:00" in md
+    assert "**趋势**：退化 — 整体质量下滑，售后场景恶化" in md
+    assert "### 指标变化" in md
+    assert "| 整窗（总体） | 80.0% → 75.0%（-5.0pp） | 100.0% → 50.0%（-50.0pp） | 120ms → 150ms（+30.0ms） |" in md
+    assert "| 售后场景 | 70.0% → 60.0%（-10.0pp） | — → —（—） | 100ms → 140ms（+40.0ms） |" in md
+    assert "### 问题演变" in md
+    assert "**[持续] 售后答非所问**（场景：售后场景）" in md
+    assert "  问题仍未收敛" in md
+    assert "**[消解] 响应偏慢**" in md
+    assert "### 建议落实情况" in md
+    assert "**[部分落实] 补充售后知识库**" in md
+    assert "  仅覆盖部分问题" in md
+    assert "**[新增] 新增建议：监控时延**" in md
+
+
+def test_render_campaign_markdown_comparison_comes_after_analysis():
+    md = render_campaign_markdown(_campaign_report(), analysis=_analysis(), comparison=_comparison())
+    assert md.index("## 智能分析") < md.index("## 周期对比")
+
+
+def test_render_campaign_markdown_without_comparison_unchanged():
+    md = render_campaign_markdown(_campaign_report())
+    assert "周期对比" not in md