Abstract: Wetlands have been paid great attention around the world for its particularity and importance in ecological environment, so nowadays wetland research has been in a hot field. Hongze Lake is rich in wetland resources and is of great significance in the regional ecoenvironment. In this paper, Hongze Lake was taken as the study area. Combining with normalized difference vegetation index (NDVI), optimized soil adjusted vegetation index (OSAVI) and the soil brightness vegetation index (BVI), greenness vegetation index (GVI), wetness vegetation index (WVI) of tasseled cap transformation. The classification has been carried out based on knowledge rules. By using classification model tailored for the classification of Hongze Lake wetlands, wetland types were effectively distinguished. Then the classification was applied to MSS and TM images of 1973, 1984 and 2006 in Hongze Lake. The results showed that the classification method based on knowledge rules is simple and easy to eliminate interference information in classification. The use of the mentioned indices in decision model can increase the overall accuracy of classification. Our study showed that the coverage of open water and wetland vegetation in Hongze Lake drastically reduced over the past 33 years from 1973 to 2006. It is necessary to take effective measurements for the protection of Hongze Lake wetlands.