Due to human activities and surface movement make a great change of land cover, it is necessary to carry out environmental monitoring of landscape dynamic and accurate mapping. A kind of landscape dynamic monitoring system based on geographic information system(GIS) is proposed. The multi-temporal decision tree classifier and morphological image processing technology are used for the system. It monitors the land use/cover change(LUCC) dynamically according to the spectrum and surface model data of remote sensing image. Experiment is conducted in local image in the Landsat 5 TM database. The results show that, this method can achieve 90.77% recognition accuracy, has lower computation time and can detect vegetation distribution(plant/Vegetation Restoration), city(city/demolition) and physical morphology(fluvial geomorphic change/slope instability).