Abstract: The high-density resistivity method is critical for detecting water hazards in mine tunnels. However, in deep and complex geological conditions, this method faces technical challenges such as the strong multi-solution nature of resistivity and insufficient precision in locating water-rich anomalies, hindering precise prevention and control of deep mine water hazards. To address these issues, this study proposes an innovative forward and reverse tripole dual-frequency induced polarization (IP) method for coal mine tunnels, leveraging parallel electrical instruments for potential acquisition and establishing a data decoding workflow. Using full-space homogeneous media and the geological characteristics of the A-group coal floor strata in the Panxie mining area (Huainan), geoelectric models with spherical and columnar water-rich anomalies were constructed for dual-frequency IP numerical simulations. By comparing the response characteristics and imaging effects of apparent complex resistivity and apparent frequency dispersion rate among the forward and reverse tripole, single-sided tripole, and quadrupole arrays, the study reveals performance differences. The results show that the quadrupole array produces simple curve shapes, facilitating anomaly location identification, but it has small anomaly amplitudes, low sensitivity to deeply buried anomalies, and poor imaging resolution. The single-sided tripole array offers larger anomaly amplitudes but is affected by array asymmetry, leading to deviations between anomaly extreme points and model cores, compromising imaging accuracy. In contrast, the forward and reverse tripole array achieves dual-way data fusion, using the 'orthogonal point' of apparent complex resistivity and the 'anti-orthogonal point' of apparent frequency dispersion rate as composite criteria. This array retains the high sensitivity of the tripole array while eliminating the systematic deviation of the single-sided tripole array, enabling accurate anomaly positioning and improving imaging quality. Field implementation demonstrates that the proposed method, through joint inversion and comprehensive interpretation of apparent complex resistivity and apparent frequency dispersion rate, significantly reduces the multi-solution nature of electrical anomalies, accurately identifies water-rich zones, and provides a novel approach for developing precise water hazard exploration systems in deep mines.