This paper presents an algorithm for retrieval of soil moisture content from hh-, hv-, and vv-polarized radar observation using a genetic algorithm with the semi-empirical polarimetric scattering model for bare soil surfaces. The semi-empirical polarimetric scattering model was developed empirically based on an extensive database comprising the polarimetric radar measurements using the ground-based scatterometers and the JPL airborne synthetic aperture radar (AirSAR) system. The input parameters of the scattering model are the volumetric soil moisture content, the rms height, the correlation length, incident tingle, and radar frequency. The outputs of the scattering model are vv-, hh-, hv-polarized backscattering coefficients its well its the phase parameters of the dearee of correlation and the co-polarized phase-difference. The inversion technique in this study is based on a genetic alaorithm. The vv-, hh- and liv-polarized backscattering Coefficients are used as the cost function of the Genetic algorithm. The outputs of the inversion model are the soil moisture and the rms surface height. One of main objectives of this study is to optimize the cost function with appropriate combination of three polarization data. The inversion model is verified by comparing the inverted results with in-situ measurement data.