Impedance　spectroscopy　is　widely　used　as　a　cable　fault　location　technique　with　high　sensitivity,　however,　due　to　interference　among　multiple　fault　points,　signal　attenuation,　and　complex　operating　environment,　there　are　some　misjudged　location　points　in　the　obtained　location　curves.　In　this　paper,　the　localization　of　cable　defects　is　achieved　by　impedance　spectroscopy　and　Fourier　transform,　and　the　location　of　fault　points　is　further　obtained　by　training　and　testing　a　series　of　localization　curves　obtained　by　SVM　and　verified　by　field　data.　The　results　show　that　SVM　can　classify　the　points　on　the　localization　curves　and　realize　the　intelligent　identification　of　fault　points,　which　reduces　the　impact　of　misjudged　peaks　in　the　impedance　spectrum,　obtains　better　localization　results,　effectively　improves　the　accuracy　of　cable　fault　location.　©　2022　IEEE.