<p>In this paper, a new kernel/window/context based approach is followed in volumetric soil moisture appraisal. After estimation of dielectric constant of the current pixel as well as neighbouring pixels, an inverse distant-weighted dielectric constant (IDW dielectric) is computed, from which soil moisture followed. Multi-Polarization Synthetic Aperture Radar image-data of NOVA SAR (S band) and EOS 4 (C band) satellites in tri pol. and quad pol. respectively have been used for the retrieval of soil moisture. A simple, yet logical, improved water cloud model has been envisaged, which is applicable to both bare fields as well as fields covered with moderate crops. The backscatter contribution from soil was estimated and the so estimated soil backscatter in like polarizations (HH and VV) was applied to Oh et al. model for deriving volumetric moisture content in soil, since vegetation will be a “noise” as far as the soil moisture application is concerned. For Quad. polarization SAR data, Oh’s model, which was prepared for tri pol. data in the past, has been augmented. Results in this research with statistical ‘t’ test, which informs whether or not a significant difference exists between the actual and computed values, indicate that the estimated soil moisture after incorporating the improved water cloud correction and IDW dielectric show better performance/match with ground soil moisture than standard per pixel calculations. Correlation coefficient was also calculated between the actual (ground) soil moisture values and the computed, with IDW dielectric method yielding 0.796 when the crops were fully grown and about 0.9 when crops were not fully mature.</p>