2. RELATED RESEARCH At present, there are few related literatures on Cryptocurrency prediction research, and most scholars' research focuses on the field of "stock" trading. Compared with Cryptocurrency, the stock exchange market developed earlier, the relevant policies were more perfect, and the reference documents were more detailed and perfect. In order to accurately judge and analyze the trend of stock price, people have developed many methods. Technicians often ignore real-world events when judging future stock market trends. To the technician, stock prices are regarded only as typical time series data with complex patterns. Through proper preprocessing and modeling, patterns can be analyzed and beneficial patterns can be extracted from them. The information used for technical analysis mainly LSTM Cryptocurrency Prediction Model based on Attention Mechanism Haorui Wan1 , Chenglin Li 1 , Kaiming Zhao1 , Ruilin Wang1 1.