The invention relates to a cooperative control method and system for chassis movement and target strike of a ground unmanned vehicle. According to the method, a built reinforcement learning parameter model is trained and tested through a built simulation scene, the trained reinforcement learning parameter model is obtained, a special vehicle type and the reinforcement learning parameter model can be organically combined, and in addition, in the actual environment, various information collected by a vehicle sensor in real time is input to serve as input of deep reinforcement learning, cooperative control over ground unmanned vehicle chassis movement and target strike is finally achieved, cooperation of an autonomous maneuvering module and an autonomous task module can be achieved, the task completion time is shortened, and the task execution effect is improved. Furthermore, the reinforcement learning method based on simulation data can enable the data acquisition cost to be low, and compared with a mathematical model method based on rules, the method can be applied to a new scene only by properly modifying input data, output actions and reward functions, and is better in universality.