Aiming at the problems of low convergence accuracy and falling into local optimal solution in solving UAV 3D path planning by traditional swarm intelligence algorithm, a UAV 3D path planning method based on improved golden monkey algorithm is proposed. Chaotic mapping is introduced to initialize the population to enhance the randomness of individuals. Nonlinear decreasing model is used to balance the global search and local exploitation ability of the algorithm, and elite reverse learning strategy is used to make the individual better jump out of the local optimum. The experimental results show that compared with other swarm intelligence algorithms, the improved golden monkey algorithm has better optimization ability.