Because path planning can provide reference instructions for multi-unmanned aerial vehicle (UAV) flight control, and current particle swarm trajectory planning algorithms have the disadvantages of slow convergence and low success rate, a multi-UAV with comprehensive improvement of particle swarm is proposed. The collaborative path planning algorithm considers UAV performance constraints, obstacle and threat constraints, space collaboration and time collaboration constraints. Firstly, through the linear adjustment of the learning factors, the balance between particle inertia and optimal behavior was achieved; secondly, chaotic initialization was introduced to improve the quality of particle distribution; then, a replacement strategy was designed based on the idea of genetic mutation, and a speed regulation mechanism was also proposed , Which improves the convergence speed of the algorithm. Finally, a comprehensive improved particle swarm algorithm is used for simulation verification. The planning results have a high success rate, fast convergence speed, and low path cost, which shows the effectiveness of the improved algorithm.