Themulti-channel SAR image domain STAP methods, with a long coherent processing time, can achieve very high performance of clutter restrain andmoving targetdetection. However, the array distortion and non-homogeneity environments have notbeen taken into consideration in the existing image domain STAPmethods. This paper studies their influence on the performance of the image domain STAP, and then proposes a new multi-channel SAR image domain STAP method that can cope with array distortion and non-homogeneity environments. The introducedmethod firstapplies the PowerSelection and Training (PST) method to select the training data, and then proposes anew Improved Sample Matrix Inversion (ISMI) method to remove themoving targets in the training data.With the training data eliminating the moving targets, the covariance matrix of the clutter is estimated.Afterwards the array distortion is estimated based on the covariance matrix, and the array distortion is calibrated later during the image domain STAP processing. The experiments based on real clutterdata prove the validity of the introducedmethod.