Aiming at the application of MIMO array radar in imaging fuze， the equivalent phase center error in near-field imaging of fuze is analyzed and the quadratic error compensation method based on Taylor expansion is given. Based on the premise of limited space on the missile， considering the influence of the coupling of the transmitting and receiving antennas， combined with the performance of the antenna beam pattern， the radius distribution interval of the transmitting and receiving array elements is given as the effective set. Based on the effective set， this paper proposes a sparse MIMO array optimization method based on two-step particle swarm optimization （TS-PSO）， and obtains the optimized two-dimensional sparse MIMO array. Based on the received echo signal， the equivalent phase center error and motion error are compensated to obtain the imaging results. The simulation results show that the two-dimensional sparse MIMO array based on TS-PSO can meet the requirements of resolution and imaging quality of fuze near-field imaging， which lays a theoretical foundation for the next practical application.