Abstract: This research was carried out with two phases of remotely sensed images covering Nanjing City in 1988 and 1998 respectively to monitor the land use change.At first,two images were normalized at the same radioactive level of 1988's TM image,registered with high geometric accuracy less than 0.5 pixel errors and purified training regions,after pretreatment of TM data.Then maximum likelihood classification is applied for both images.Finally the GRID model of Arc/Info software is used to algebraically calculate two classified maps with AML codes and obtained information of land use change.Through analyzing the result indicates that in the past period of ten years arable land decreased largely was mainly due to urban extension,and forest increased was due to some protective policies and activities of forestation.