Abstract: Multiple-input and multiple-output (MIMO) radar is a new radar system developed internationally in recent years. Because of the superiority of the system, MIMO radar has better performance in many aspects such as target detection, angel estimation, target recognition and high resolution imaging due to the use of space diversity and signal diversity technology, compared with the traditional radar system. In order to reduce the array dimension and hardware complexity of wideband MIMO radar system which exists in the two dimensional high resolution imaging in common, and make the application of MIMO radar imaging technology become true, a nonlinear array model is proposed, combined with the improved back-projection(BP) time domain imaging algorithm. The proposed nonlinear array model can acquire the array element intervals by the cross-range resolution requirement of imaging system and the choice of the imaging scene, and so can reduce evidently the array dimension and hardware complexity of imaging system. Meanwhile, the application of time-domain imaging algorithm can avoid the limitation of sampling theory which exists in the frequency-domain imaging algorithm, make the array design become flexible, and furthermore reduce the array dimension and hardware complexity of imaging system. At last, simulation experiments show the correctness and effectiveness of this nonlinear array model.