In the airborne synthetic aperture radar system, the estimation accuracy of Doppler parameters affects the image quality directly. This paper estimates Doppler centroid using the averaged azimuth spectrum, and it estimates the Doppler frequency rate by computing the shift between sub-images. The raw data are windowed to reduce the affection of various noises in the system, and improve the estimation accuracy. Results of the algorithm can be used to compensate the phase error caused by motion error of the airborne. The method is finally verified with a set of actual experimental data.