Transient electromagnetic method (TEM) surveys in engineering geophysics research can be used to detect geological anomalies caused by water. In this study, a hybrid-dimensional TEM data inversion method was developed to produce high-resolution results for detecting water hazards in coal mines, solving the problem of misestimating water hazard anomalies. This method utilizes 3-D inversion with a reliable initial model to accurately map the subsurface resistivity and identify the distribution of water hazards in coal mines. The initial model is created using weighted lateral constraint inversion (LCI), which is suitable for sedimentary formations. Furthermore, when unstructured tetrahedral meshes are used for model discretization, the enhanced 3-D inversion method can effectively mitigate the impact of high-relief topography and enhance the positioning accuracy of anomalous bodies. The application of the synthetic model demonstrates that hybrid-dimensional inversion can accurately restore underground electrical distributions. The field data processing shows the successful detection of water hazards in a goaf in a coal mine in Ordos, Inner Mongolia, China. The resistivity distributions of sedimentary strata can be precisely restored using the new method, which fits well with the known geology and underground drilling.