This paper presents an improved algorithm for downscaling SMAP soil moisture using C-band SAR data. It is developed using multi-temporal C-band Radarsat Constellation Mission (RCM) compact polarimetric (CP) SAR data. In this improved algorithm, the effect of vegetation on soil moisture retrieval from SAR data is minimized by using a normalized scattering based empirical model, in which vegetation contribution is quantified using the volume scattering derived from RCM CP decomposition. The influence of soil surface roughness is eliminated by using multi-temporal data. The multi-temporal SMAP soil moisture and RCM CP data (simulated from Radarsat-2 QuadPol data) are the only inputs of this downscaling model. The model is tested in southern Ontario, Canada to downscale 36 km resolution SMAP soil moisture to 1 km. The downscaled results have good agreement with the in-situ soil moisture collected in June of 2017 with an unbiased root-mean-square-error (RMSE) of 0.047 m3/m3 and a coefficient of determination (R2) of 0.43. The results suggest that the improved algorithm can be applied for C-band RCM CP data to provide continuous soil moisture mapping over large area at higher resolutions because of RCM's high-revisit frequency and large areal coverage characteristics.