Remote sense geoanomaly is one kind of important information for m etallogenic prognosis. But the oreforming possibility of particular remote s ense geoanomalies may be very different. An expert system method for oreform ing probability appraisal is proposed in this paper. The inference procedure from the three types of remote sense geoanomalies (i.e. image parameter value anomalies, lineaments and circulars structure anomalies and multisource data fusion anomalies) to oreforming geoanomalies is outlined. Detailed discussion on the system structure, middle layer knowledge representation and characteristics of oreforming anomalies are carried out.