This paper analyzes the current network security situation in the era of big data. At the same time, the author summarizes the theory of network security management from the aspects of risk control and security assessment. On this theoretical basis, a network security management platform based on artificial intelligence big data is constructed. The platform has built a complete network security situational awareness and early warning architecture that integrates data collection and sorting, data storage, offline analysis and discovery, real-time correlation detection, threat warning and situation presentation, and supports the whole process of security event processing. The experimental results prove that the multi-model ensemble algorithm used in the proposed system combines the advantages of each model, making the overall accuracy rate up to 99.24%.