diff --git a/TODO_NEXT_STEPS.txt b/TODO_NEXT_STEPS.txt
index 5802ea8..59272ba 100644
--- a/TODO_NEXT_STEPS.txt
+++ b/TODO_NEXT_STEPS.txt
@@ -4,7 +4,8 @@
 ================================================================================
 
 【已完成】
-   - V10.320 補市場情報 candidate queue review AI summary Telegram dispatch report input：新增 read-only report input builder、獨立 report route extension、UI 按鈕與 deployment readiness smoke target，在 archive summary 後整理 report input sections、report contract、message evidence 與 dispatch audit traceability；同步把 Telegram HTML 送出前的 `<br>` 統一轉為換行，避免 Bot API 回 400。API/UI 不讀 approval/Telegram token、不呼叫 LLM、不派送 Telegram、不開 DB、不寫檔、不產報表、不更新 review_state、不掛 scheduler。
+   - V10.321 修正 Telegram HTML 發送格式：所有 `sendMessage` / `sendPhoto` caption 在 HTML parse mode 送出前會把 `<br>` / `<br/>` / `<BR />` 統一轉成換行，避免 Telegram Bot API 回 `Unsupported start tag "br"` 造成告警或報告送出失敗。
+   - V10.320 補市場情報 candidate queue review AI summary Telegram dispatch report input：新增 read-only report input builder、獨立 report route extension、UI 按鈕與 deployment readiness smoke target，在 archive summary 後整理 report input sections、report contract、message evidence 與 dispatch audit traceability；API/UI 不讀 approval/Telegram token、不呼叫 LLM、不派送 Telegram、不開 DB、不寫檔、不產報表、不更新 review_state、不掛 scheduler。
    - V10.319 補市場情報 candidate queue review AI summary Telegram dispatch archive summary：新增 read-only archive summary builder、POST endpoint、UI 按鈕與 deployment readiness smoke target，在 Telegram dispatch archive 後整理 message identity、dispatch audit、artifact manifest 與後續 report input sections；API/UI 不讀 approval/Telegram token、不呼叫 LLM、不派送 Telegram、不開 DB、不寫檔、不更新 review_state、不掛 scheduler。
    - V10.318 收緊 Elephant Alpha HITL 告警治理：`ea_escalation` 只有真正含 SKU/價格比較的 actions 才排成 TOP 待審 SKU 卡片；非 SKU 診斷改為「待確認事項」，並用測試鎖住價格類低信心但無 DB/Hermes 實證時只 suppress、不寫 human_review、不發 Telegram，避免空泛告警打擾人工審核。
    - V10.317 修正 PChome 比價覆蓋率分子：`fetch_competitor_coverage()` 的 valid_matches 改成 `ACTIVE + 有 MOMO 最新價` 商品與有效 PChome `identity_v2` 價格的交集，不再把非活躍或無 MOMO 現價的舊 competitor_prices 列入覆蓋率，避免 daily/growth/PPT/AI 報表高估比價資料品質。
diff --git a/config.py b/config.py
index 82123a8..24c8f79 100644
--- a/config.py
+++ b/config.py
@@ -320,7 +320,7 @@ YOUTUBE_API_KEY = os.getenv('YOUTUBE_API_KEY', '')
 # ==========================================
 # 系統版本與路徑
 # ==========================================
-SYSTEM_VERSION = "V10.320"
+SYSTEM_VERSION = "V10.321"
 LOG_FILE_PATH = os.path.join(BASE_DIR, 'logs/system.log')
 public_url = PUBLIC_URL  # 用於模板顯示
 
diff --git a/docs/AI_INTELLIGENCE_MODULE_SOT.md b/docs/AI_INTELLIGENCE_MODULE_SOT.md
index 08586b3..106d797 100644
--- a/docs/AI_INTELLIGENCE_MODULE_SOT.md
+++ b/docs/AI_INTELLIGENCE_MODULE_SOT.md
@@ -610,3 +610,4 @@ POSTGRES_HOST=momo-db
 | 2026-05-20 | 指定日期競品簡報可能混用目前 `competitor_prices` 快取價 | V10.315 起 `fetch_competitor_comparison_results()` 有 start/end date 時改用 `competitor_price_history` 期間快照，MOMO 價格取報表結束日前最新價；即時報表才使用目前有效 `competitor_prices` |
 | 2026-05-20 | PChome 覆蓋率分子可能被非活躍或無 MOMO 現價 SKU 膨脹 | V10.317 起 `fetch_competitor_coverage()` 的 `valid_matches` 改為 active MOMO latest price 與有效 PChome `identity_v2` 價格交集，確保 daily/growth/PPT/AI 看到的比價資料品質不被舊快取列高估 |
 | 2026-05-20 | EA HITL 告警可能把非 SKU 診斷誤排成待審 SKU，或在缺少 DB/Hermes 實證時打擾人工 | V10.318 起 `ea_escalation` 僅對含 SKU/價格比較的 actions 使用競價卡片；非 SKU 診斷改為「待確認事項」。價格類低信心事件若無 DB/Hermes 實證，測試鎖定只 suppress、不寫 human_review、不發 Telegram |
+| 2026-05-20 | Telegram HTML parse mode 不支援 `<br>`，可能導致告警或報告送出 400 | V10.321 起 Telegram template 發送前會把 `<br>` / `<br/>` / `<BR />` 轉為換行；保留其他 HTML 標籤，非 HTML parse mode 不改寫 |
diff --git a/services/telegram_templates.py b/services/telegram_templates.py
index 740bc59..a4484a9 100644
--- a/services/telegram_templates.py
+++ b/services/telegram_templates.py
@@ -30,6 +30,7 @@ sys_log = logging.getLogger("TelegramTpl")
 
 TELEGRAM_BOT_TOKEN_ENV = "TELEGRAM_BOT_TOKEN"
 TELEGRAM_CHAT_IDS_ENV  = "TELEGRAM_CHAT_IDS"
+_TELEGRAM_HTML_BR_RE = re.compile(r"<\s*br\s*/?\s*>", re.IGNORECASE)
 
 
 # ══════════════════════════════════════════════════════════════════════════════
@@ -55,6 +56,14 @@ def _get_chat_ids() -> list:
         return []
 
 
+def _sanitize_telegram_html(text: str, parse_mode: Optional[str] = "HTML") -> str:
+    """Telegram HTML 不支援 <br>，統一轉為換行避免 sendMessage 400。"""
+    value = str(text or "")
+    if parse_mode and str(parse_mode).upper() == "HTML":
+        return _TELEGRAM_HTML_BR_RE.sub("\n", value)
+    return value
+
+
 def _send_telegram_raw(text: str, chat_ids: Optional[list] = None,
                        reply_markup: Optional[Dict[str, Any]] = None,
                        parse_mode: Optional[str] = "HTML") -> bool:
@@ -69,7 +78,7 @@ def _send_telegram_raw(text: str, chat_ids: Optional[list] = None,
         chat_ids = [-1003940688311]  # fallback
 
     url = f"https://api.telegram.org/bot{token}/sendMessage"
-    payload = {"chat_id": chat_ids[0], "text": text}
+    payload = {"chat_id": chat_ids[0], "text": _sanitize_telegram_html(text, parse_mode)}
     if parse_mode:
         payload["parse_mode"] = parse_mode
     if reply_markup:
@@ -106,7 +115,7 @@ def send_telegram_with_result(text: str, chat_ids: Optional[list] = None,
     errors: List[str] = []
 
     for chat_id in chat_ids:
-        payload = {"chat_id": chat_id, "text": text}
+        payload = {"chat_id": chat_id, "text": _sanitize_telegram_html(text, parse_mode)}
         if parse_mode:
             payload["parse_mode"] = parse_mode
         if reply_markup:
@@ -151,7 +160,7 @@ def send_photo(photo_bytes: bytes, caption: str = "",
     try:
         r = requests.post(
             url,
-            data={"chat_id": chat_ids[0], "caption": caption[:1024],
+            data={"chat_id": chat_ids[0], "caption": _sanitize_telegram_html(caption, parse_mode)[:1024],
                   "parse_mode": parse_mode},
             files={"photo": ("chart.png", photo_bytes, "image/png")},
             timeout=30,
diff --git a/tests/test_telegram_triaged_alert_format.py b/tests/test_telegram_triaged_alert_format.py
index 219dcd3..c7c2ad6 100644
--- a/tests/test_telegram_triaged_alert_format.py
+++ b/tests/test_telegram_triaged_alert_format.py
@@ -1,4 +1,41 @@
-from services.telegram_templates import triaged_alert
+from services import telegram_templates
+from services.telegram_templates import _sanitize_telegram_html, triaged_alert
+
+
+def test_telegram_html_sanitizer_converts_br_tags_to_newlines():
+    msg = _sanitize_telegram_html("第一行<br>第二行<br/>第三行<BR />第四行")
+
+    assert "<br" not in msg.lower()
+    assert msg == "第一行\n第二行\n第三行\n第四行"
+    assert _sanitize_telegram_html("第一行<br>第二行", parse_mode=None) == "第一行<br>第二行"
+
+
+def test_send_telegram_with_result_sanitizes_html_payload(monkeypatch):
+    sent_payloads = []
+
+    class Response:
+        ok = True
+        status_code = 200
+        text = "ok"
+
+    def fake_post(url, json=None, timeout=None):
+        sent_payloads.append({"url": url, "json": json, "timeout": timeout})
+        return Response()
+
+    monkeypatch.setattr(telegram_templates, "_get_bot_token", lambda: "telegram-token")
+    monkeypatch.setattr("requests.post", fake_post)
+
+    result = telegram_templates.send_telegram_with_result(
+        "第一行<br>第二行<br/>第三行",
+        chat_ids=[101, 202],
+        parse_mode="HTML",
+    )
+
+    assert result["ok"] is True
+    assert result["sent"] == 2
+    assert [item["json"]["chat_id"] for item in sent_payloads] == [101, 202]
+    assert all(item["json"]["text"] == "第一行\n第二行\n第三行" for item in sent_payloads)
+    assert all(item["json"]["parse_mode"] == "HTML" for item in sent_payloads)
 
 
 def test_ea_escalation_uses_structured_incident_brief():