The　fault　location　of　high-voltage　direct　current　(HVDC)　transmission　line　usually　needs　to　determine　the　characteristic　quantity　and　then　calculate　it.　In　this　paper,　the　artificial　extraction　process　is　removed　and　the　end-to-end　parallel　neural　network　is　applied　for　fault　location.　The　multiport　flexible　DC　transmission　system　model　is　built　on　the　MATLAB/Simulink　simulation　platform,　the　line　fault　current　under　different　fault　conditions　is　extracted,　and　the　CNN-LSTM　parallel　neural　network　model　is　built.　With　the　proposed　fault　distance　measurement　method　and　its　corresponding　parallel　neural　network,　average　error　less　than　0.2　km　can　be　achieved　under　a　large　amount　of　test　case.　Compared　with　single　CNN　and　single　LSTM　networks,　this　method　maintains　high　accuracy　and　reliability,　which　can　handle　the　fault　resistances　range　from　0.01Ω　to　200Ω.　The　simulation　results　validate　the　feasibility　of　proposed　fault　location　method.　©　2021　IEEE