Abstract: Normalized difference vegetation index (NDVI) of vegetation cover in the study area is computed based on TM images using remote sensing (RS) and geographic information system (GIS) software. NDVI is used as a weight number to calculate a weighed average of theoretical forest biomass potential on different forest land uses in each administrative district.Then all weighed average data as raster values in the study area is applied to generate a GIS layer entitled spatial density distribution (SDD) of theoretical biomass. A GIS layer of available forest biomass potential in the study area is derived from the SDD layer of theoretical biomass. Finally,the layer of available forest biomass potential is applied to analyze site-selection for both collection sites and processing sites to make use of forest biomass and optimize them.In the process of qualitative site-selection analysis,on the basis of constrained factors derived from economic and operational feasibility,GIS with its powerful functions of spatial analysis is adopted to determine candidate zones where the sites should be situated.This paper proposes a quantitatively analytical procedure and an optimization method referenced location-allocation modeling and service-area modeling.Assessment scores of the candidate zones (centroids) are computed through the method and they are applied to setting priority.Finally,this paper theoretically gives optimal sites where collection sites and processing sites may be located,including supply areas around both existing and potential processing sites.