diff --git a/app/Article.py b/app/Article.py
index b4df54c..bfef595 100644
--- a/app/Article.py
+++ b/app/Article.py
@@ -39,9 +39,9 @@ def get_today_article(user_word_list, had_read_articles):
             "index" : 0,  # 为 article_ids 的索引
             "article_ids": []  # 之前显示文章的id列表，越后越新
         }
-    if had_read_articles["index"] > len(had_read_articles["article_ids"])-1:
+    if had_read_articles["index"] > len(had_read_articles["article_ids"])-1:  # 生成新的文章，因此查找所有的文章
         rq.instructions("SELECT * FROM article")
-    else:
+    else:  # 生成阅读过的文章，因此查询指定 article_id 的文章
         rq.instructions('SELECT * FROM article WHERE article_id=%d' % (had_read_articles["article_ids"][had_read_articles["index"]]))
     rq.do()
     result = rq.get_results()
@@ -53,23 +53,33 @@ def get_today_article(user_word_list, had_read_articles):
     d3 = get_difficulty_level(d1, d2)
 
     d = None
+    result_of_generate_article = "not found"
     d_user = load_freq_history(user_word_list)
     user_level = user_difficulty_level(d_user, d3)  # more consideration as user's behaviour is dynamic. Time factor should be considered.
     text_level = 0
-    if had_read_articles["index"] > len(had_read_articles["article_ids"])-1:  # 下一篇
-        flag_get_article = False
-        for reading in result:
-            text_level = text_difficulty_level(reading['text'], d3)
-            factor = random.gauss(0.8,
-                                  0.1)  # a number drawn from Gaussian distribution with a mean of 0.8 and a stand deviation of 1
-            if reading['article_id'] not in had_read_articles["article_ids"] and within_range(text_level, user_level, (8.0 - user_level) * factor):  # 新的文章之前没有出现过且符合一定范围的水平
-                d = reading
-                had_read_articles["article_ids"].append(d['article_id'])  # 列表添加新的文章id；下面进行
-                flag_get_article = True
-                break
-        if not flag_get_article:
+    if had_read_articles["index"] > len(had_read_articles["article_ids"])-1:  # 生成新的文章
+        amount_of_had_read_articles = len(had_read_articles["article_ids"])
+        amount_of_existing_articles = result.__len__()
+        if amount_of_had_read_articles == amount_of_existing_articles:  # 如果当前阅读过的文章的数量 == 存在的文章的数量，即所有的书本都阅读过了
+            result_of_generate_article = "had read all articles"
+        else:
+            test_count = 0
+            for k in range(3):  # 最多尝试3次
+                test_count += 1
+                for reading in result:
+                    text_level = text_difficulty_level(reading['text'], d3)
+                    factor = random.gauss(0.8, 0.1)  # a number drawn from Gaussian distribution with a mean of 0.8 and a stand deviation of 1
+                    if reading['article_id'] not in had_read_articles["article_ids"] and within_range(text_level, user_level, (8.0 - user_level) * factor):  # 新的文章之前没有出现过且符合一定范围的水平
+                        d = reading
+                        had_read_articles["article_ids"].append(d['article_id'])  # 列表添加新的文章id；下面进行
+                        result_of_generate_article = "found"
+                        break
+                if result_of_generate_article == "found":  # 用于成功找到文章后及时退出外层循环
+                    break
+            print(test_count)
+        if result_of_generate_article != "found":  # 阅读完所有文章，或者循环3次没有找到适合的文章，则对 index+=1 进行回滚
             had_read_articles["index"] -= 1
-    else:  # 上一篇
+    else:  # 生成已经阅读过的文章
         d = random.choice(result)
         text_level = text_difficulty_level(d['text'], d3)
 
@@ -86,7 +96,7 @@ def get_today_article(user_word_list, had_read_articles):
             "answer": get_answer_part(d['question'])
         }
 
-    return had_read_articles, today_article
+    return had_read_articles, today_article, result_of_generate_article
 
 
 def load_freq_history(path):
diff --git a/app/templates/userpage_get.html b/app/templates/userpage_get.html
index 6264ef0..e2cb9d4 100644
--- a/app/templates/userpage_get.html
+++ b/app/templates/userpage_get.html
@@ -71,10 +71,14 @@
                 <button onclick="toggle_visibility('answer');">ANSWER</button>
                 <div id="answer" style="display:none;">{{ today_article['answer'] }}</div><br/>
             </div>
-        {% else %}
+        {% elif result_of_generate_article == "not found" %}
             <div class="alert alert-success" role="alert">
                 <p class="text-muted"><span class="badge bg-success">Notes:</span><br>No article is currently available for you. You can try again a few times or mark new words in the passage to improve your level.</p>
             </div>
+        {% elif result_of_generate_article == "had read all articles" %}
+            <div class="alert alert-success" role="alert">
+                <p class="text-muted"><span class="badge bg-success">Notes:</span><br>You've read all the articles.</p>
+            </div>
         {% endif %}
     </div>
 
diff --git a/app/user_service.py b/app/user_service.py
index 712fd13..a306724 100644
--- a/app/user_service.py
+++ b/app/user_service.py
@@ -135,7 +135,7 @@ def userpage(username):
         words = ''
         for x in lst3:
             words += x[0] + ' '
-        had_read_articles, today_article = get_today_article(user_freq_record, session.get('had_read_articles'))
+        had_read_articles, today_article, result_of_generate_article = get_today_article(user_freq_record, session.get('had_read_articles'))
         session['had_read_articles'] = had_read_articles
         if today_article is None:
             session["found_article"] = False
@@ -147,6 +147,7 @@ def userpage(username):
                                session=session,
                                # flashed_messages=get_flashed_messages(), 仅有删除单词的时候使用到flash，而删除单词是异步执行，这里的信息提示是同步执行，所以就没有存在的必要了
                                today_article=today_article,
+                               result_of_generate_article=result_of_generate_article,
                                d_len=len(d),
                                lst3=lst3,
                                yml=Yaml.yml,