The presented spatio-temporal modeling based on long time series from 12 satellite images provides considerable experience for processing NDVI in the framework of identification of land-cover classes and also, to a certain degree, cropland variability with its impact on food security.