To explore the prediction effect of network security situational awareness on network vulnerabilities and attacks under the background of big data, this study constructs a predictive index system based on the network security situational awareness model. Based on the improved cuckoo algorithm, the cuckoo search radial basis function neural network is used to predict the situation. The weight value in the model is determined by the hierarchical analysis method, vulnerability simulation is conducted by Nessus software and network attack simulation is conducted by Snort software, and then the situation is evaluated by a fuzzy comprehensive evaluation method. Finally, Jquery and Bootstrap software is used to develop the system. The results show that the cuckoo search radial basis function model proposed in this study could predict network security situations more accurately than the radial basis function model, cuckoo search back-propagation neural network model, genetic algorithm radial basis function model and Support vector machine model based on particle swarm optimization model.