Abstract: Multiple cropping system (double- and triple-crop in one year) is an important agricultural system in China, which is significant to increase grain yield and promote agricultural economic development. Multiple Cropping Index (ratio of sown area and cropland area) is fluctuating year by year due to the changing natural conditions and rural social status, so, getting the change information timely is undoubtedly very important to assess grain yield change and its reasons, as well as to make scientific decision for agricultural development. China is a vast country with complex cropping system, however, the traditional statistic method cannot meet the government demands on cropping system variation. Satellite remote sensing is the most effective means to identify land cover pattern and change on a regional scale, so it should be an important way to discriminate multiple cropping characteristics on a large scale. In this study, the authors attempt to extract multiple cropping information from 8km 10-day composite AVHRR/NDVI time series images according to the phenological metrics and farmland practices (such as sowing and harvest) temporal features, so as to research the cropping system change in the past twenty years. Official agriculture statistics contain the information about sown areas and cropland areas of each county, which provide a useful data source for the validation of remotely sensing based crop rotation retrieval. Multiple cropping system spatial pattern was mapped using this method, then its availability and precision were discussed.