Abstract: Poyang Lake is the largest fresh lake and lies in red soil hilly region of China. Poyang basin is an important agricultural production base, especially paddy field and an area suffered from serious flood disaster. In this study, a new idea of classification for the clustering of Poyang basin land cover has been proposed based on the cover spectral response and characteristics of seasonal variability by using multi-temporal SPOT4-VEGETATION sensor data (NDVI) while land cover type is named after color. Then a two-level land cover classification system has tentatively been upbuilt. Based on it, the preliminary land cover map of Poyang basin has been derived from the unsupervised classification of NDVI digital images by applying ISODATA algorithm in PCI software and Spatial Analyst Function in ArcMap. According to the area, the total land cover density reaches 94.97%: evergreen cover is 40.05%, crop-forest-grassland mixed cover 38.29%, crops 16.63% in proportion to the Poyang basin at present.