Abstract: Radiometric calibration is a powerful means to quantify the SAR （Synthetic Aperture Radar） system performance. Compared with that before launching, there is a considerable distortion of the antenna elevation pattern after the spaceborne SAR is in orbit, which will debase the quality of SAR images greatly. An accurate measurement of the antenna elevation pattern is an important aspect of spaceborne SAR radiometric calibration. The external calibration of distributed targets based on the uniform backscattering is an optimal choice to measure the double-way antenna elevation pattern. Amazon rainforest is the most stable, most uniform and flattest distribute external target in the world. With a backscattering almost independent of the incidence, Amazon rainforest is regarded as an ideal radiometric calibration site. However, because the SAR data sets collected by our own spaceborne SAR are still unavailable at present, the estimation of unknown parameters of SAR antenna pattern model based on the simulated spaceborne SAR power image has been investigated in this paper. A successful computer simulation of how to determine the antenna elevation pattern in orbit by simulated Amazon rainforest SAR power images has been made, which is significant for the radiometric calibration of spaceborne SAR in our country.