diff --git a/lib/utils/wechat-mp.js b/lib/utils/wechat-mp.js
index 8e3278019..96723029c 100644
--- a/lib/utils/wechat-mp.js
+++ b/lib/utils/wechat-mp.js
@@ -35,6 +35,18 @@ const replaceTag = ($, oldTag, newTagName) => {
     oldTag.replaceWith(NewTag);
 };
 
+const detectOriginalArticleUrl = ($) => {
+    // No article content get, try the original url
+    if (!$('#js_content').text()) {
+        return $('#js_share_source').attr('data-url');
+    }
+    // Article content is too short, try the first link
+    if ($('#js_content').text().length < 80) {
+        return $('#js_content a').attr('href');
+    }
+    return null;
+};
+
 /**
  * Articles from WeChat MP have weird formats, this function is used to fix them.
  *
@@ -147,7 +159,16 @@ const fetchArticle = async (ctx, url, bypassHostCheck = false) => {
         const author = $('meta[name=author]').attr('content');
         let summary = $('meta[name=description]').attr('content');
         summary = summary !== title ? summary : '';
-        const description = fixArticleContent($('div#js_content.rich_media_content'));
+        let description = fixArticleContent($('div#js_content.rich_media_content'));
+
+        // No article get or article is too short, try the original url
+        const originalUrl = detectOriginalArticleUrl($);
+        if (originalUrl) {
+            // try to fetch the description from the original article
+            const originalResponse = await got(normalizeUrl(originalUrl, bypassHostCheck));
+            const original$ = cheerio.load(originalResponse.data);
+            description += fixArticleContent(original$('#js_content'));
+        }
 
         let pubDate;
         const publish_time_script = $('script[nonce][type="text/javascript"]:contains("var ct")').first().html();