According to the fault characteristics of fault line and non-fault line, a method of fault line selection for small current grounding system based on fuzzy entropy and discrete Fréchet distance was proposed in this paper. A complete set of empirical mode decomposition (CEEMDAN) was used to decompose the transient zero-sequence current of each line, and the characteristic IMF component was selected according to kurtosis and continuous mean square error. Then, the correlation analysis of the characteristic IMF component of each line was carried out to construct a similarity characteristic matrix and determine whether it was a line fault or a bus fault. If the fault is a line fault, fuzzy information is granulated on the characteristic IMF components of each line. Then, the three groups of Low, R, and Up after granulation of each line were mixed and superimposed. And the discrete Fréchet distance were calculated, combined with the membership function, the fuzzy entropy with complex influence factors was obtained, and a complex feature matrix was constructed. The fault selection was carried out by using the dual characteristic matrix. After a lot of simulation analysis, it is proved that the accuracy of the proposed fault selection method is not affected by the transition resistance and fault closing angle, and the noise resistance is strong.