As space exploration grows, so does the amount of space junk, even affecting the normal work of spacecraft and satellites. An improved particle swarm optimization(PSO) algorithm applied to perform space junk recovery tasks in parallel with multi-spacecraft is proposed in this paper. First, adaptive inertia weights and acceleration coefficient are designed to prevent the proposed algorithm from falling into local optimum; Second, utilizing the idea of variation and crossover in genetic algorithm for processing particle matrices can improve the efficiency of the algorithm. Finally, the simulation results show that the proposed algorithm provides reasonable assignment results, the operation time is reduced by 3%, and the revenue is improved by 13 % at least.