For a given incidence angle at the snow surface, a greater snow density causes a greater change in the incidence angle at the snow-ground interface; for a given snow density, however, a larger incidence angle at the snow surface results in a greater change in the refractive angle in the snow layer, by comparing the difference of incidence angle at the snow-ground interface and the air-snow interface with different snow density. Algorithm for estimating dry snow density used backscattering measurements with polarimetric SAR at L-band frequency is developed based on simulation of the surface backscattering components ghh,and gvv using the IEM model and regression analysis. The comparison of the estimated snow density from SAR L-band images with that from field measurements during the SIR-C/X-SAR overpass shows root means square error of 0.050 g/cm3. It shows that this algorithm can be accurately used to estimate dry snow density distribution. For a given incidence angle at the snow surface, a greater snow density causes a greater change in the incidence angle at the snow-ground interface; for a given snow density, however, a greater incidence angle at the snow surface results in a greater change in the refractive angle in the snow layer, by comparing the difference of incidence angle at the snow-ground interface and the air-snow interface with different snow density. Algorithm for estimating dry snow density used backscattering measurements with polarimetric SAR at L-band frequency is developed based on simulation of the surface backscattering components ghh, and gvv using the IEM model and regression analysis. The comparison of the estimated snow density from SAR L-band images with that from field measurements during the SIR-C / X-SAR overpass It shows that this algorithm can be accurately used to estimate dry snow density distribution.