Aiming at the challenge of ambiguous boundary identification of anomalous bodies in coal fields using the transient electromagnetic method under complex geological conditions, this study proposes the introduction of the DBSCAN clustering algorithm to determine the boundaries of target bodies in mine transient electromagnetic detection. Taking three typical concealed disaster-causing bodies in mines—low-resistivity goafs, high-resistivity layered coal seam dislocation faults, and composite anomalous bodies such as coal seam collapse columns and water-filled roadways—as research objects, the response characteristics are obtained through transient electromagnetic forward modeling and apparent resistivity imaging. The DBSCAN algorithm is then applied to cluster analyze the apparent resistivity data, achieving precise identification of anomalous body boundaries. By selecting the neighborhood radius based on the inflection point in the k-distance graph and establishing clustering criteria, the algorithm effectively enhances the contrast between anomalous bodies and the background field. This significantly reduces the boundary positioning error of typical anomalous bodies such as water-rich goafs and greatly improves the identification accuracy of complex geological bodies. Finally, transient electromagnetic detection data from the 5130 working face of the Yanjiahe Coal Mine are selected for anomalous boundary identification. The results demonstrate that the DBSCAN clustering analysis algorithm can effectively identify anomalies in transient electromagnetic imaging results, clearly determining anomalous boundaries. This provides a new approach for the precise identification of water-rich goaf ranges and other disaster-causing body boundaries.