With the extensive use of drones in both military and civilian fields, multi-drone collaborative trajectory planning has become a research hotspot. This paper proposes a multi-drone three-dimensional trajectory planning method based on a multi-population gray wolf optimization algorithm (MP-GWO), aimed at addressing multi-objective optimization problems in complex environments. Building on the GWO algorithm, this study first introduces a multi-population collaborative mechanism, dividing the search space into multiple sub-populations through clustering for separate optimization, thereby enhancing global search capability. Furthermore, based on a greedy approach and random initialization method, the average fitness value of the initial population is improved. Finally, the convergence factor in the GWO algorithm is adjusted from linear to a nonlinear control parameter; the adjusted convergence factor decreases rapidly in early iterations to enhance global search capability, avoiding local optima, and subsequently decreases gradually in later iterations to increase local search capability. Experimental results indicate that compared with the standard GWO algorithm, the MP-GWO algorithm demonstrates significant improvements in both convergence speed and trajectory quality, effectively planning three-dimensional collaborative trajectories that meet constraint conditions.