Ground moving target indication (GMTI) is an important research field in multichannel-synthetic aperture radar systems (SAR). The robust principal component analysis (RPCA) method can separate the sparse matrix of moving targets from the low-rank matrix of static backgrounds in image or video data. As the correlation coefficient variation caused by phase different between channels is weak for SAR, it results in that the conditions of applying RPCA method cannot be satisfied. To solve this problem, an improved moving target detection method based on RPCA is proposed in this paper. By analysis, the phase different between channels is transformed to the shifting in time-frequency domain. Therefore, the correlation coefficient is decrease between channels, which means that the sparse matrix of moving targets is extracted with higher probability. Simulation and analysis illustrate the effectiveness of the proposed method eventually.