Water bursting in mines remains one of the major disasters that threatens safe coal production. The traditional detection techniques of water-rich regions in mines can no longer meet new safe production standards and requirements. To solve the constraints on the selection of traditional finite-difference time-domain (FDTD) step length by the Courant-Friedrichs-Lewy stability condition, the unconditional stability Crank-Nicolson FDTD (CN-FDTD) algorithm was introduced into the detection of underground total space, aiming to further increase calculation speed and detection accuracy. The boundary condition of the perfectly matched layer (PML) and numerical stability condition for mine detection were analyzed by studying the basic principle of CN-FDTD. Meanwhile, a detection model of underground water-rich regions was constructed. Influences of excitation function, frequency of excitation source, distance between the anomalous body and the interface, and size of the anomalous body on water-rich region detection were analyzed through Wavenology EM and MATLAB simulation by using the forward algorithm. Results demonstrate that given the same conditions, the Ricker wavelet excitation source has better detection effect than the Blackman-Harris Window (BHW) Function, Gaussian and Delta excitation sources. When the anomalous body is 70 m away from the interface, the excitation source with a central frequency of 15 MHz is the best detection frequency. In addition, forward modeling reveals that when the anomalous body is 50 m away from the interface, the echo signals of a 3 cubic metres anomalous body could still be detected clearly. Theoretically, the existence of an anomalous body can be determined through this echo signal, and its location can be determined through imaging. The study verifies the effective application of CN-FDTD in water-rich region detection in mines. In addition, CN-FDTD solves the constraints of traditional FDTD by stability condition and increases the accuracy of water-rich region detection in mines. This study provides certain theoretical references for the accurate detection of water-rich anomalous bodies in mines.