Abstract: Aiming at the problems of unclear distribution of water accumulation hidden dangers, insufficient resolution of traditional detection means and others in the goaf of third mining area in Shouyang Luyang Xiangsheng Coal Industry, based on electromagnetic induction theory, transient electromagnetic method is applied to carry out identification and research on large-scale, high-density water rich abnormal bodies, combined with the spatial structure of regional faults and collapse column dense areas, a transient electromagnetic survey network covering 2.16 km² is set up, and a multi-channel synchronous acquisition system is adopted to obtain 4003 sets of response data; Through the exponential decay model fitting and apparent resistivity inversion, a two-dimensional and main control layer slice map is constructed to achieve spatial precise positioning of water bearing abnormal bodies and goafs, five main low resistance abnormal areas are successfully identified, the classification of water rich levels is clear, among them, the resistivity of level I area is lower than 22 Ω·m. By comparing with the downhole water inflow data and the result of drilling pumping tests, the anastomosis rate reaches 87.2%. The research verifies the effectiveness of shallow layer interference suppression technology, especially in the imaging accuracy of 0-50 m shallow layer water richness zones. By introducing wide frequency domain induced field enhancement technology, the anomaly recognition rate of shallow layers is increased from 68% to 92%. The research results provide quantitative support and spatial identification basis for proactive prevention and control of coal mine water damage, prediction of water sources for downhole mining and excavation replacement, and disaster evaluation.