In the context of the new era, the development strategies of countries and research institutions around the world are gradually expanding to the ocean. As a hot technology, unmanned ship technology has been widely used in military and civil fields. Compared with traditional ships, it has great development potential in terms of safety and economy. The emergence of unmanned ships has contributed to China's maritime security. The collision avoidance path planning capability of the unmanned ship is one of its key technologies, and it is also an important manifestation of its artificial intelligence. Based on artificial intelligence algorithm, this paper studies the maritime safety control system of unmanned ships. The system completes the control system design based on the position and distance of obstacles, water leakage detection, own position and attitude, image acquisition and remote monitoring. Afterwards, the design experiments were carried out to test, and the results showed that after continuous evolution, the path with higher fitness function was retained, and the optimal path was created in the 100th generation, indicating that the system designed in this paper can effectively provide the optimal path for unmanned ships. airway. This platform is small in size and flexible in use, and has complete basic functions, which can provide a useful design reference for ensuring the safety of unmanned ships. In this paper, an effective control system is designed by introducing artificial intelligence algorithms into the field of unmanned ship maritime security.