Automatic mapping models for the evaluation of mineral - resource potentials are usually applied to statistical integration of multi-genetic geological information in order to automatically delineate metallogenetic -prospecting areas based on the GIS technology. In this paper, a new automatic mapping model for mineral resource potentials assessment based on the GIS technology is proposed, named the combined metallogenetic confidence model, developed based on the certainty theory that is a reasoning model with uncertainty. With this model, a raster map of the metallogenetic confidence can be created on the basis of a favorable or an unfavorable metallogenetic indicator pattern. Then, according to the special combining rule of metallogenetic confidence, all raster maps can be statistically integrated into the raster map of combined metallogenetic confidence. Based on this raster map, areas with relatively high values of combined metallogenetic confidence are delineated as the metallogenetic prospecting targets. And a contour map of combined metallogenetic confidence also can be generated based on this raster map. The model has been applied to predict multi - metallogenetic prospecting targets in the district from Duolanasayi to Ashele in northern Xinjiang. The predicted targets by the model are compared with those by the weighted evidence model. The two predicted results are very similar to each other. This proves that this new model is effective and practical.