Abstract: Four-dimensional reconstruction of urban areas and man-made infrastructure is one of the most important applications of differential SAR (Synthetic Aperture Radar) tomography. In this situation, the urgent problem of differential SAR tomography is how to achieve satisfying elevation-velocity resolution using sparse samples of space-time plane and preserve azimuth-range resolution simultaneously. Compressive sensing is a perfect theory to resolve this urgent problem. First, we prove the applicability of compressive sensing for differential SAR tomography theoretically. Then, the resolution ability of differential SAR tomography based on compressive sensing is discussed using simulation experiment with the real system parameters. Finally，we apply this imaging method to the real data of Envisat-ASAR. Both the theoretical analysis and satisfying results of Envisat-ASAR data processing confirmed the applicability of this differential SAR tomograhic imaging method.