Abstract: In view of the problem that the data retrieval by traditional searching methods based on keywords and topic words cannot satisfy the real needs that emerge in the geological data research, this study provides a solution for semantic level retrieval. A geological data ontology model related to mineral resource predication is built through a detailed expression of the concepts, characteristics, relationship and instances. The geological data linked retrieval system is established. By associating the geological data ontology, the experiment in this research proves that the proposed ontology shows a significant improvement in data discovery. The linked retrieval system could obtain the geological data in local area network and research relatively in semantic level to meet the requirement of big data application service. The application of geological data ontology in this paper will have an important significance for the accurate data discovery, data tapping and data sharing in the era of geological big data.