By remote sensing technique, a wealth of mineralization-related information is acquired from the research in remote-sensing geological characteristics of uranium minerals from Huashan granite body of Guangxi. The correlation between geological unit, geological structure, geological complexity, alterative, hot abnormal information and the mineralization is analyzed by MapGIS. Ten types of metallogenic evidence factors are selected, and the GIS-based evidence weighted model of metallogenic prediction is established. The weights of various factors via the unit cells are calculated based on the analysis of various metallogenic factors from the evidence weighted model. Four levels prospective regions are fixed with the posterior probabilities got from the total information in each unit cell.