The　existing　fault　detection　and　identification　methods　for　series　arc　current　are　based　on　the　feature　extraction　of　signal　singularities　for　arc　current　with　zero　moment,　mutation　rate　of　climb　and　harmonic　component,　etcval　However,　there　is　an　erroneous　judgement　which　is　caused　by　the　current　signal　singularities　of　nonlinear　load　in　rated　duration,　inductive　load　in　startup　process,　etc.　in　distribution　line.　According　to　the　principle　of　load　side　voltage　not　to　be　impacted　by　the　current　signal　singularities　in　normal　operation　for　distribution　circuit,　a　novel　arc　fault　detection　and　identification　method　for　load　side　voltage　is　put　forward,　which　can　not　only　obtain　series　arc　fault　feature,　but　also　avoid　the　defection　of　existing　arc　fault　current　method.　By　arc　fault　identification　decision-making　functions　modeling　with　morphological　wavelet,　the　combination　identification　model　of　morphology　open-filter　with　the　fourth　dimension　wavelet　transform　is　selected　to　analyze　arc　fault　experiment　detection　signal　of　six　kinds　load　side　voltage.　Finally,　the　fourth　dimension　wavelet　component　criterion　thresholds　of　arc　fault　signal　for　load　side　voltage　are　provided,　that　are　ten　times　the　normal　condition　for　kinds　of　load.