Despite the capability of remote sensing to direct observation of soil moisture content, the radiances measured by sensors are usually affected by different soil and atmosphere parameters. Therefore, understanding the importance of selecting the optimal features for soil moisture recognition, the application of fuzzy logic to perform intelligent feature selection is a distinguished line of research. In the following, the selected features were used in two widely used classifiers (SVM (Support Vector Machine) and MLP (Multi-Layers Perceptron) artificial neural network) in order to soil moisture classification. These classifiers were found competitive with the best available machine learning algorithms. In other words, the main purpose of this model is to select the least number of features based on fuzzy logic aligning with increasing the accuracy of soil moisture classification. The proposed method was applied and validated using observations carried out for the Iran region. In order to compare the soil moisture classification accuracy using the features selected by fuzzy-based model, a different scenario was also considered. In the latter case, vegetation cover (NDVI), soil surface temperature (LST), and topography as selected features for soil moisture classification, were entered into the above-mentioned classifiers. The reason for choosing these three features among all the features is their significant effect on the amount of soil moisture. The results obtained were very encouraging and indicated about 8% improvement on soil moisture classification accuracy using the proposed feature selection method.