Abstract: A high-performance Medical Resource Database System (MRDS) is described. Based on ontology, the system implements the unification of vocabularies in the domain of medical image and makes up for the semantic gaps between original data and the information inquired out by customers, medical textual information and image information and the resource database system and other systems. By semantic inference, inquiring navigation and processing the correlation feedback information inquired out, the system can understands the intention of customer better. The efficiency of inquiry is thus improved by means of paralleled computing. Experiment results showed that the system is able to unify the high efficiency, high recall ratio and high precision ratio into one.