This work presents the fuzzy-rough spatial model for the real-time decision support systems for complex dynamic systems. The proposed model uses the dynamic fuzzy-rough topology and the three-level spatial hierarchy, which represents discontinuous spatial areas at the upper level, continuous regions at the middle level and their approximation by the grid of cubic cells at the lower level. The boundaries of regions are blurred using fuzzy or rough sets of cells. The implementation of the model based on six GIS layers is presented. It reduces the computational complexity and provides the acceptable performance of DSS.