Based on the CBERS1/CCD data, GIS and GPS techniques, and combined with conventional data and field investigation, digital image processing and classifying methods are applied to interpret resources of landuse, water and forest in Guixi city, Jiangxi Province. Some GPS locating data are chosen to interpret image, others remaining to check up image-classified results, which improved the objectivity and efficiency of interpreting image. By use of section observational data in subtropical mountainous regions and locating data for seventeen years in Guixi mountain region, twelve models are built to describe changes of climate factors with heights. On the basis of interpolating digital counter onto each 20-meter grid, the climate resources in the grid are calculated. Consequently, seven kinds of crop's comprehensive climate division are given by means of logical arithmetic of GIS space analysis models. Also the results of divisions are analyzed.