Abstract: Mineral resources constitute a critical strategic foundation for supporting the national economy and social development. To enhance the efficiency of mineral resource exploration, particularly in challenging areas such as those with thick Quaternary cover, permafrost regions, coastal zones, and deep geological environments where traditional methods are often difficult to deploy, artificial intelligence technologies, including big data, deep learning, and knowledge graphs, have been extensively adopted. This application has substantially boosted mineral exploration prediction efficiency, offering pivotal technological support for the intelligent transformation of the field. Intelligent methods for mineral exploration prediction can synergistically integrate multi-source heterogeneous data, e.g., geological, geophysical, geochemical, remote sensing, and historical exploration data with computer technologies like geoscience knowledge graphs and deep learning. This integration facilitates the automation and intellectualization of mineral exploration, thereby significantly improving overall efficiency, and it has consequently become a major research hotspot. This paper systematically reviewed the development trajectory of current intelligent methods for mineral exploration prediction. The research and application of these methods in both two-dimensional mineral exploration and three-dimensional deep mineral exploration prediction were analyzed, focusing on four key technical aspects: machine learning, deep learning, knowledge graphs, and three-dimensional geological modeling. Furthermore, the primary challenges facing these intelligent methods were outlined, and their future research trends were discussed. The development of large language models by computers will undoubtedly provide technical support for the next generation of intelligent methods for mineral exploration prediction.