Soil moisture remains one of the most important indices influencing soil resistance to the shear forces of dropping and flowing water or blowing wind. Physically based models as well as experiments at various spatial and time scales confirm that firmly. There is however a major obstacle in wide utilization of soil moisture information in wide-area erosion intensity monitoring - the availability of reliable data sources. The research described in that paper is aimed at finding a reliable methodology for the estimation of current moisture of the topsoil, which can be than used to estimate potential or actual erosion rates at given erosive factors’ intensity and duration. Thermal satellite imagery was chosen as source of data for soil humidity. Algorithms were developed for combining scenes from different satellites were developed aiming at achieving maximum temporal and spatial resolution of the output data. A NDTI model for transforming the thermal band images into soil humidity rasters were developed basing upon thermal inertial model. The models were validated with manual on ground measurements of soil humidity using high accuracy TDR soil moisture meter, revealing relatively good fit of the developed models with correlation coefficient at the level of R2=0,41. A simple wind erosion model was chosen to demonstrate the usage of the soil moisture data overlaid with soil and land use maps. However the method of estimating the temporal soil moisture has a bigger application potential, eg. in single event soil erosion models, plant water deficit estimations and defense.