In the current era of the rapid development of computer networks, data security is of paramount importance. This research aims to construct a computer network data security encryption system based on artificial intelligence. During the research process, first, the advantages and disadvantages of traditional encryption algorithms were delved into in-depth, and then the integration of neural networks and genetic algorithms was introduced. Through the training of a large number of network data samples, the system can automatically learn data features and achieve more precise encryption strategies. In terms of encryption algorithms, the improved AES (Advanced Encryption Standard) algorithm is utilized, combined with the genetic algorithm to optimize the key-generation process, enhancing the complexity and security of keys. After testing, the data encryption efficiency of this system has increased by 30% compared with traditional encryption systems, effectively reducing the risk of data being cracked and significantly enhancing the security of computer network data.