With the continuous development of the Internet in recent years and the continuous expansion of the user base, ransomware attacks have become more frequent, the technologies used have become more diversified, and the purpose of the attack is no longer satisfied with just encrypting the user’s files or locking the user’s system, but stealing the user’s data while encrypting to achieve secondary ransom. How to detect and identify ransomware has become a top priority for Internet security and a hot topic in cyberspace security. This paper uses the 1DCNN+LSTM model with a self-attention mechanism to perform detection and verification on a self-built data set. The experimental results show that the 1DCNN+LSTM algorithm using the self-attention mechanism achieved an accuracy of 97.22% on the self-made ransomware network traffic data set.