Polarimetric soil moisture retrieval is among the main objectives of leading synthetic-aperture-radar satellite missions since it allows to systematically analyze costly-to-obtain polarization information to increase retrieval accuracy. In this letter, we present the results of modeling the L-band entropy (H) and alpha (��) values as a function of soil dielectric constant and roughness using a second-order small-perturbation model to simulate the polarimetric soil backscattering. Modeling results are then compared to bare soil uninhabited aerial vehicle synthetic aperture radar (UAVSAR) data acquired simultaneously to in situ field campaigns in Canada during SMAPVEx12. Our model is able to correctly predict observed ranges of H and �� and to consistently model dielectric constant and roughness changes. Nevertheless, a systematic overestimation of �� is observed when compared with the analyzed UAVSAR data set. Taking UAVSAR data as benchmark, theoretical reasons for this mismatch are analyzed.