This paper presents an inverse model for SAR backscatter X - band designed to snow and ice surfaces. The model inversion enables estimating the morphological characteristics of snow cover from SAR data. X Cosmo-Skymed sensor data was used in the experiment tested here for the region of Union Glacier - Antarctica. We observed a precision in determining the estimated size of snow grains about 94%. This result depicts the applicability of the analyzed model as an alternative way for extracting information from X-SAR data