diff --git a/yt_dlp/extractor/niconico.py b/yt_dlp/extractor/niconico.py
index 5b3a7e45ff..fa18858b30 100644
--- a/yt_dlp/extractor/niconico.py
+++ b/yt_dlp/extractor/niconico.py
@@ -2,6 +2,7 @@ import datetime as dt
 import functools
 import itertools
 import json
+import re
 
 from .common import InfoExtractor, SearchInfoExtractor
 from ..networking.exceptions import HTTPError
@@ -720,21 +721,27 @@ class NicovideoSearchBaseIE(InfoExtractor):
     def _entries(self, url, item_id, query=None, note='Downloading page %(page)s'):
         query = query or {}
         pages = [query['page']] if 'page' in query else itertools.count(1)
+        is_old = False
         for page_num in pages:
             query['page'] = str(page_num)
             webpage = self._download_webpage(url, item_id, query=query, note=note % {'page': page_num})
 
-            server_response = self._search_regex(r'<meta\s+name=["\']server-response["\']\s+content=["\']([^"\']+)["\']', webpage, 'videos json')
-            sr_json = self._parse_json(unescapeHTML(server_response), 'search response') or {}
-            results = traverse_obj(
-                sr_json,
-                ('data', 'response', '$getSearchVideoV2', 'data', 'items', ..., ('id')),
-            ) or []
+            results = re.findall(r'(?<=data-video-id=)["\']?(?P<videoid>.*?)(?=["\'])', webpage)
+            has_next = None
+            if results:
+                is_old = True
+            if not is_old:
+                server_response = self._search_regex(r'<meta\s+name=["\']server-response["\']\s+content=["\']([^"\']+)["\']', webpage, 'videos json', fatal=False)
+                sr_json = self._parse_json(unescapeHTML(server_response) or '{}', 'search response') or {}
+                results = traverse_obj(
+                    sr_json,
+                    ('data', 'response', '$getSearchVideoV2', 'data', 'items', ..., ('id')),
+                ) or []
+                has_next = traverse_obj(sr_json, ('data', 'response', '$getSearchVideoV2', 'data', 'hasNext'), default=None)
 
             for item in results:
                 yield self.url_result(f'https://www.nicovideo.jp/watch/{item}', 'Niconico', item)
 
-            has_next = traverse_obj(sr_json, ('data', 'response', '$getSearchVideoV2', 'data', 'hasNext'), default=None)
             if has_next is not None and has_next is False:
                 break
             if not results: