With the increasing complexity of the network environment and the continuous evolution of security threats, the traditional network security protection and evaluation methods are facing more and more challenges. It is urgent to find a solution to improve the protection efficiency and response speed. Artificial intelligence (AI) technology, especially machine learning and deep learning, is widely regarded as a powerful tool to enhance network security because of its excellent ability in data analysis and pattern recognition. This study aims to improve the accuracy of Intrusion Detection System (IDS) and malware identification by implementing a series of machine learning algorithms. Specific methods include using deep learning network to analyse and classify network traffic data, and using natural language processing technology to extract effective information from network threat intelligence. In addition, this paper also explores risk assessment models optimized by machine learning, which can dynamically analyse and evaluate potential security vulnerabilities. All participants have made remarkable progress during their study, and even their performance in the phishing recognition course has improved by nearly 42%. This study confirms the application value of AI algorithm in computer network information security protection and evaluation, which can not only greatly improve the performance of existing systems, but also bring new research and application directions to the field of network security.