In order to improve the security of the secure wireless network system, this paper studies the security of the wireless network system based on the deep reinforcement learning network. Moreover, this paper divides the entire network into a domain structure, and each domain structure includes ordinary nodes, council nodes, key nodes, and dynamic gateway nodes. The key node selects the dynamic gateway node in the domain, and after the dynamic gateway node processes the message, it is forwarded to the gateway node in the adjacent domain. In addition, this paper studies the application of multi-objective optimization in trust management mechanism in secure routing. Furthermore, this paper builds a secure wireless network system based on the deep reinforcement learning network based on the deep reinforcement learning network, and builds a system architecture. Finally, this paper tests the secure wireless network system based on deep reinforcement learning network through simulation experiments. The research results show that the safety factor of the system is relatively high.