It is proposed the forecasting model for the situation in space image processing systems. The analysis of the currently existing methods and models for forecasting time series is carried out, their advantages and disadvantages are revealed. The model for forecasting the situation based on satellite images has been improved. It is an adaptive sampling model. As a selection criterion, the minimum sum of squares of deviations for previous forecasts is used. As a result, it ensures the reliability and accuracy of the forecast of the situation, and on the basis of this forecast it is possible to make effective decisions.