In view of the background of high-intensity confrontation game between the enemy and ourselves in combat, this paper studies the evaluation of node importance based on maneuvering path and the construction of confrontation game model based on battlefield environment, and constructs a game path planning method based on improved neural network. The method is mainly divided into two modules: pre-path planning and real-time path planning. On the basis of pre-path planning, the initial maneuvering path scheme is obtained, and according to the new battlefield environment information and the maneuvering position information of the missile unit, the gain of the neural network is continuously updated until the end of the maneuvering path planning, and the real-time path planning is realized. The real-time path planning has an important reference value for the path selection of normal navigation maneuver.