Abstract: Three-dimensional (3D) geological modeling is regarded as an effective technical method for locating deep-seated minerals. However, its application in deep metallogenic prediction of sandstone-type uranium deposits remains limited. Focusing on the Hadatu-Saihan Gaobi area in the Erlian Basin, this study developed a 3D geological model for deep metallogenic prediction by integrating geological, geophysical, and remote sensing data. Given the characteristics of multivariate geoscience information, this study proposed a layered 3D implicit modeling method. Specifically, for modeling at depths less than 1 000 m, geological and drilling data, along with ground electromagnetic survey results, were primarily used. In contrast, for modeling at depths exceeding 1 000 m, the results from 3D joint gravity and magnetic inversion were utilized. The resulting 3D geological model reveals that primary strata in the study area include the Neogene-Paleogene, Lower Cretaceous, Permian, Carboniferous, and Neoproterozoic strata, with prominent rock masses comprising granites and intermediate-basic rocks. The elevated and slightly elevated fields of aeroradiometric uranium content around the known uranium deposit are associated with the migration, deposition, and enrichment of uranium-bearing materials, as well as fault-related tectonic movements. Through three-dimensional metallogenic prediction based on metallogenic condition analysis, three metallogenic prospect areas with geological characteristics similar to the known uranium deposit were identified. This study provides a novel philosophy for the interpretation of aeroradiometric data and the exploration of deep uranium deposits in basins.