Soil moisture is an important parameter for hydrological and climatic investiga- tions. It also plays a critical role in the prediction of erosion, flood or drought. In this paper, we explore the use of the support vector machine technique for modeling soil moisture inversion. LS-SVM is improved by the standard SVM and has more attractive properties. Experimental tests are carried by using the different set of training and test data. The methodologies have been applied to two sets of data to retrieve soil moisture and obtained the root mean squared error (RMSE) and the determination coefficient (R2 ). The emissivity model (Q/H model) is applied to acquire the brightness temperature. The frequencies of interest include 1.4 GHz (L-band) of the soil moisture and ocean salinity (SMOS) sensor at two incidence angles and 6.9 GHz (C-band) of the advanced microwave scanning radiometer (AMSR) viewing angle of 55 deg. The effectiveness is assessed by considering various combinations of the input features. This study demonstrates the great potential of LS-SVM in the retrial of soil moisture from passive microwave remotely sensed data.