A dual-channel airborne synthetic aperture radar (SAR) imaging approach is proposed to acquire well-focused images of ground moving targets (GMTs) after compensating for the platform trajectory deviation (PTD). The phase error attributed to either PTD or target motion is given by a polynomial of slow time. The SAR image before motion compensation is segmented into multiple subimages, and a compressive-sensing (CS) technique is applied on each subimage to estimate the polynomial coefficients featuring PTD. A range-frequency reversal transform (RFRT) and fractional Fourier transform (FrFT) are combined to estimate the motion parameters and position of each moving target. The proposed approach is verified with real data and simulated data, respectively, and the moving targets in the acquired image are well recognized.