Airborne differential synthetic aperture radar interferometry (D-InSAR) is a kind of potential technology to survey the surface deformation. In this paper, the errors of two usually used modes of this technology, namely, two-pass and three-pass approaches, are compared and analyzed. In the analysis, all the errors are decomposed and sorted according to the principle of independence, and the coupling between motion error and other errors is considered. Based on the analysis, analytical expressions are derived for the deformation measurement errors in the two modes. The results demonstrate that when the amplitude of the motion error is small, the three-pass approach can reduce the requirement on the precision of the external topography data, so it distinctly outperforms the two-pass approach; when the amplitude of the motion error is large, high-precision topography data are needed in the three-pass approach, and as a result, its advantage over the two-pass approach is not distinct any more.