The stock market is one of the most important part of the financial market. However, it is difficult for investors to extract information from data. This paper studied the analysis of the announcement during the suspension period. It offers feasible suggestions, hopefully. It established a prediction model based on the content of announcements. The prediction is about the stock price. It analyzes the feasibility of using text classification to predict the trend of stock after the resumption. The prediction process contains the following three steps. First, it collects the announcement of the stock during the suspension period, Second, it processes these texts by nature language processing techniques. These techniques include Chinese word segmentation, feature word extraction, text representation. In this step, the announcements are transformed into text vectors composed by a specific weight. Third, it uses classification algorithms to generate prediction models. In this step, different stock index category might use different classification algorithm. This paper focused on the Chinese stock market.