In　a　small-current　grounding　system,　the　pole-to-ground　fault　may　cause　the　voltage　drop　in　the　fault　pole　and　the　voltage　rise　in　the　other　poles.　In　flexible　DC　distribution　systems,　Severe　voltage　variation　may　shorten　the　insulation　lifetime　of　the　equipment,　which　leads　to　great　concerns　on　the　safety　issue.　In　addition,　the　existence　of　high　transition　resistance　degrades　the　accuracy　of　fault　detection　methods,　thus　further　affecting　the　reliability　of　the　system.　Therefore,　it　is　essential　to　explore　an　advanced　technology　for　faulty　feeder　detection.　This　study　proposes　a　faulty　feeder　detection　method　based　on　the　characteristics　of　transient　zero-mode　current　(TZMC)　in　multi-frequency　bands.　The　change　rate　of　zero-mode　voltage　is　applied　as　the　protection　activation　criterion.　Then,　the　characteristic　matrix　is　constructed　via　computing　the　fuzzy　entropy　of　TZMC　in　each　frequency　band.　Finally,　the　faulty　feeder　can　be　identified　by　conducting　fuzzy　C-means　on　the　characteristic　matrix.　This　proposed　method　is　tested　through　simulations　on　the　PSCAD/EMTDC　platform,　which　successfully　demonstrates　its　outstanding　adaptability,　reliability,　and　accuracy.