Abstract: The mutual couplings between the conductors and sheath of the distribution cable are complex. After a single-phase grounding fault occurs in the cable, the steady features are not obvious and the transient features are variable in complex conditions (high-impedance faults, arcs, etc.), making fault cable location difficult in a complex distribution network. Aiming at the structural features of the typical distribution cable, the feasibility of decoupling the signals in time domains is firstly verified. The general method for calculating the phase-mode transformation matrices is proposed. Then, according to the states of the moduli before and after the fault, the featured moduli, effectively reflecting the fault, are selected. Finally, considering the topology of the distribution cable network and measurement deployment, the fault section location criterion based on the space-measurement matrix is created by analyzing the amplitude distribution of the frequency component in the moduli and the polarity coefficients of the current moduli at the branch node. A 10kV distribution cable fault simulation model is created by using PSCAD/EMTDC. The calculation results of the typical cases demonstrate that the faults can be accurately located by the proposed method under different conditions, which indicates its advantages over the current methods.