The ability to estimate the soil moisture content on a global scale from dual-frequency (C- and X-bands) microwave AMSR-E data is investigated in this paper. The C-band AMSR-E data were calibrated by comparing AMSR-E data with past measurements of SSM/I and SMMR sensors over some reference targets as the Amazon forest and the Ocean. An algorithm for the retrieval of soil moisture, based on a simplified form of the radiative transfer theory, was tested and subsequently inverted by using the Nelder-Mead method. The algorithm, which used both the brightness temperature at C-band and the polarization index at X-band, was successfully validated on data collected on two areas in Italy and US, during the MASMEx02 and the SMEX02 experiments, respectively