Abstract: Cropland bio-energy intensive application is an important way of solving energy and environmental problems in China. Since crop residues are not distributed centrally and continuously, the intensive application of cropland bio-energy is different from that of the traditional energy, i.e. coal, oil, natural gas, etc. Therefore, the studies of bio-energy quantity, distribution characteristics and the optimization of bio-energy intensive application are very important to help the intensive application of cropland bio-energy and the selection of optimal locations of power plants. A case study in Guangdong province, China, this paper provided a framework to quickly estimate the quantity of available cropland biomass energy and analyze its distribution pattern based on NPP model, and divide the primary collection regions by Thiessen polygon in several scales, and use genetic algorithms to optimize selecting the locations of cropland bio-energy intensive application. The results show that genetic algorithms and GIS model can solve the question of searching spatial demand point from polygon support area, and the MAUP can affect the results of GA model. When Thiessen polygon was built as primary collection regions by proximal-tolerance of 10 km, the GA model could get the best fitness values. The model can provide the effective spatial optimal method for cropland bio-energy intensive application.