diff --git a/wechat_ipad/providers/server_864/login.py b/wechat_ipad/providers/server_864/login.py
index fda8843..c30da17 100644
--- a/wechat_ipad/providers/server_864/login.py
+++ b/wechat_ipad/providers/server_864/login.py
@@ -1,4 +1,5 @@
 import qrcode
+from urllib.parse import parse_qs, urlparse
 
 from wechat_ipad.providers.server_864.base import Server864APIClientBase
 
@@ -31,12 +32,24 @@ class LoginMixin(Server864APIClientBase):
             json_body={"Proxy": proxy_value, "Check": False},
             timeout=30,
         )
-        uuid = self._pick_first(data, "UUID", "Uuid", "uuid") or ""
-        qr_url = (
-            self._pick_first(data, "QrUrl", "QRUrl", "qrUrl")
+        qr_code_url = (
+            self._pick_first(data, "QrCodeUrl", "QRCodeUrl", "qrCodeUrl")
+            or self._pick_first(data, "QrUrl", "QRUrl", "qrUrl")
             or self._pick_first(self._pick_first(data, "Qrcode", "QrCode", "qrcode") or {}, "Src", "src")
             or ""
         )
+        uuid = self._pick_first(data, "UUID", "Uuid", "uuid") or ""
+        if not uuid and qr_code_url:
+            # 864 当前真实返回的是“二维码图片服务地址”，
+            # 其中真正的扫码链接藏在 `data=http://weixin.qq.com/x/<uuid>` 这个 query 里：
+            # 1. Dashboard 需要 uuid 才能稳定生成扫码地址与展示文案；
+            # 2. 因此这里把 query 中的真实扫码链接反解出来，兼容当前 server 的返回格式；
+            # 3. 若未来某些 864 版本直接返回 UUID，本逻辑仍会优先使用原始字段，不会互相冲突。
+            parsed_qs = parse_qs(urlparse(str(qr_code_url)).query)
+            scan_data = str((parsed_qs.get("data") or [""])[0] or "")
+            if "/x/" in scan_data:
+                uuid = scan_data.rsplit("/x/", 1)[-1].strip()
+        qr_url = str(qr_code_url)
 
         if print_qr and uuid:
             qr = qrcode.QRCode(