Abstract: This paper proposes a kind of space-mapping particle swarm optimization by combining space-mapping with particle swarm optimization(PSO). In this algorithm, accurate-mesh simulation is regarded as accurate model, and coarse-mesh simulation is regarded as coarse model. Estimation of accurate result is gotten by Kalman filter mapping relation and coarse model, which can determine whether the particle is efficient for updating PSO or not, before calculating the fitness of each particle in PSO. This, in turn, can avoid the simulation of inefficient particle at some level, directly leading to reducing the computation of PSO. Simulations of E-kind patch antenna and cavity slot antenna proof the efficiency of the algorithm.