Changes in properties of land surface are caused by manmade and natural phenomena including emergencies. Fast detection of the changes is possible with the near-real time processing of satellite data only. Identification of the territory, where the changes occurred, is performed by comparison of expected and observed surface reflectances. Obtaining the reliable estimates of the expected reflectances in accordance with previous satellite observations of the region land surface for future identification is a key element for the task. The goal of this research is development and verification of the technology for expected (forecast) reflectances of the land surface from the data of MODIS spectroradiometer, installed on Terra and Aqua platforms. An algorithm for predicted reflectancies calculation based on bidirectional reflectance distribution function model from MODIS operational algorithm is discussed. A comparison of predicted and observed reflectances in first seven MODIS channels is presented. Our analysis shows that predicted reflectances from 1, 3, 4 MODIS channels are useful for the detection of pixels, where land surface properties were changed.