Applying GIS technology, surface air temperature data of 80 meteorological stati ons in Heilongjiang Province, 16 agricultural and forestry stations in Sanjiang Plain, and DEM data, a statistical model of thermal resource indices with latitu de, longitude and elevation as independent variables was constructed. The trend surface analysis and liner interpolation methods were used to calcula te and correct the thermal resource indices values. The results show that latitu de and elevation factors have significant effect on the thermal resource distrib ution, and longitude factor merely affects air temperature of warm season and ac cumulated temperature ≥0℃ and ≥10℃. Thermal resources are generally rich in the plain areas, and poor in the western and southern mountainous areas.