Abstract: This paper proposed a railway network security situational awareness model based on Convolutional Neural Network Long Short Term Memory Network (CNN-LSTM) to address the problems of low detection rate and insufficient real-time monitoring capability in current railway network security situational awareness. This model utilized the spatial feature extraction capability of CNN and combined the time series modeling advantages of LSTM to implement accurate output of anomaly detection values through full sample training. The paper selected 9 key indicators for comprehensive evaluation from three dimensions: anomaly detection performance, situational awareness effectiveness, and system availability. The experimental results show that the model has an accuracy of 97.2% in detecting abnormal attacks, while maintaining a low false alarm rate of 2.5%. This study provides an effective technical solution to enhance the awareness of railway network security situation.