Induction of attack characteristics in the network, real-time capture of situational changes caused by risk factors, detection of potential attacks and intrusions, and warning and prevention of them. This paper proposes a network situation assessment model based on Long-Short-Term Memory (LSTM) algorithm and Genetic Algorithm (GA), aiming to utilize the advantages of deep learning and evolutionary algorithms to increase the scenario assessment regarding network security’s effectiveness and precision. This paper proposes the application of LSTM algorithm in network data sequence modeling to capture temporal features and potential threat behaviors in network traffic data. At the same time, we apply the GA algorithm to optimize network security policies, searching for the optimal security policy parameters through evolutionary algorithms to improve the security and robustness of the network. The experimental results showed that on dataset 2, GA-LSTM also performed the best with an accuracy of 0.82, slightly higher than the accuracy of LSTM and GA-SVM (both 0.77).