SAR raw data is disturbed by additive system noise which influences the SAR image intensity, however, Traditional radiometric correction methods usually ignore the effects of system noise. Particularly in the case of ScanSAR, system noise is added to its own scalloping effect. In low backscatter area, radiometric correction without considering system noise will remain a residual noise which will deteriorate the quality of image. In this paper, we first quantize the effect of system noise in ScanSAR radiometric correction, and then introduce a method to estimate system noise and reduce it. Eventually, we propose a method to correct scalloping in ScanSAR system based on reduction of system noise. It reduces noise and correct antenna pattern in amplitude image. Simulations and application of the algorithm in real ScanSAR data processing show that scalloping pattern is removed and image quality is improved.