Abstract: In order to better solve multi-objective problem and improve the diversity and convergence of multi-objective optimization algorithms, an improved multi-objective particle swarm optimization algorithm is proposed. The algorithm divides the population into several subpopulations for optimization search and improves the particle velocity updating formula to expend the coverage of Pareto optimal solution set, and use inverse trigonometric logistic mapping to initialize the population to make the distribution of initial population more uniform. The time-varying variation method is used to change the external files to avoid local optimization. By comparing the performance of improved algorithm, standard multi-objective particle swarm optimization (MOPSO) algorithms and NSGA-Ⅱ on the standard test function, the effectiveness of the improved algorithm proposed in this paper is verified in an optimal radar distribution station.