Abstract: This study proposes a fault line selection method for mine power network based on wavelet decomposition waveform area matrix and grey-relational analysis (GRA) to address the insufficient utilization of fault characteristics misjudgment based on single criterion in the existing fault line selection methods for mine high-voltage power network. Specifically, the transient zero-sequence current of each line was decomposed by coif5 wavelet and the reconstruction coefficient area matrix was constructed. The low-frequency band and high-frequency band components in the area matrix were eliminated, and the characteristic scale and the preliminary result of line selection were obtained by the set line selection margin formula; the grey correlation matrix was constructed by using the zero-sequence current component of each line under the characteristic scale of line selection, and the result of line selection was again obtained by setting the line selection margin formula; the final line selection result was obtained according to equal weight voting. The proposed algorithm was verified by various simulation experiments. Results show that the line selection algorithm exhibits high accuracy and high reliability, and shows strong adaptability to different fault initial phase angles, different transition resistances, different fault occurrence distances and noise interference.