In order to solve the problem of large trajectory dispersion when the missile adopts attitude control midcourse guidance, a design method of midcourse guidance scheme trajectory based on BP neural network is proposed. A large number of attitude commands of the level flight phase under different disturbances are obtained by mathematical simulation of the position control system, and the complex mapping relationship between the disturbance input and the level flight attitude command output is learned from a large number of sample data, and the mapping relationship is deployed into the actual aircraft trajectory program. The results show that the scheme trajectory can effectively avoid external interference and greatly improve the accuracy of the midcourse guidance scheme trajectory of this type of missile without increasing the cost.