A Subsection Mean Method (SMM) for relation analysis of land surface temperature field (LSTF) and the faults is proposed. The method assumes that the center of fault zone is known. According to the distance from the fault zone, a certain region is equally partitioned. Average LST and distance in each equallydidivded zone are calculated, and then the data are plotted in the correlation diagram of the mean LST verse the mean distance. In the case study of JiangshanShaoxing fault between Jinhua and Quzhou of Zhejiang Province, this paper describes the relations between LSTF and deep large faults based on the images of the Landsat 7 ETM and thermal infrared remote sense. The results indicate that LSTF is distinctly related to the JiangshanShaoxing fault, ChangshanLizhu fault and QiuchuanXiaoshan fault and regionally, shows a stable change feature. Within the area effected by each fault, the closer the distance to the fault, the higher is the LST; The NWtrending Chun'anWenzhou fault and the EWtrending QuzhouTiantai fault are situated in the transitional zone of LST, while areas on both sides of the two faults may belong to different LSFTs.