Abstract: 3D mineral resource potential assessment is an important tool for deep ore prediction. In recent years, with the integration of machine learning algorithms, the accuracy of 3D mineral resource potential assessment has significantly improved, making outstanding contributions to the field of mineral resource exploration. The current mining elevation of the Weishan REE deposit is approximately -200m. The exploration results for deep-seated ore prospecting are not satisfactory, and further exploration is needed. In this study, 3D geological modeling was utilized to depict the spatial distribution of ore bodies, rock formations, and structures, and various geological factors were quantitatively extracted. Subsequently, the correlation between various geological factors and mineralization was analyzed using the weight of evidence, Based on this, a random forest model was established for deep ore prediction. Combining with the regional ore-forming background, two exploration targets were identified, providing insights for deep ore exploration in the Weishan REE deposit.