Wavelet　decomposition　method　is　used　to　analysis　the　vibration　signal　of　low　voltage　circuit　breaker　mechanical　properties.　According　to　the　electric　operating　mechanism　and　circuit　breaker　closing　action　sequence　relations,　drive　motor　current　signal　as　a　time　stamp　is　applied　to　effectively　extract　switching　vibration　signal.　Then,　vibration　signal　feature　vectors　are　structured　by　means　of　the　wavelet　packet　energy　spectrum,　and　BP　neural　network　are　applied　to　establish　the　three-phase　closing　asynchronous　fault　identification　model.　The　experiment　and　simulation　results　show　that　the　combination　of　wavelet　packet　energy　spectrum　and　neural　network　can　effectively　analysis　of　low　voltage　circuit　breaker　closing　synchronism.　©　2012　IEEE.