Multiple input multiple output (MIMO) synthetic aperture radar (SAR) shows much potential compared with traditional SAR in many interesting applications, but effective orthogonal waveforms design is still a challenge. Sub-chirp combination waveform has large time-bandwidth product, low cross-correlation, and low peak-average ratio, and it has been used in some MIMO SAR research. However, only the reduction of peak-average ratio is considered in the existing sub-chirp combination waveform optimal design research. In this paper, imaging performance is considered for waveform design, using cost function based on MIMO SAR general ambiguity function (GAF). Furthermore, genetic algorithm is used to obtain optimal waveforms. Finally, the optimal waveforms by the proposed method are compared with the waveform examples in the existing literature, and the simulation results show the effectivity of the proposed optimal design method.