The　key　point　to　capture　the　breakdown　information　is　how　to　improve　the　signal-to-noise　ratio　of　the　vibration　signal　for　the　fault　diagnosis　of　bearing.　However,　the　methods　of　traditional　Hilbert　demodulation,　the　traditional　generalized　filtering　demodulation　and　etc.　do　not　have　the　ability　to　improve　it.　Theoretical　analysis　and　experimental　results　show　that　it　is　difficult　to　diagnose　the　bearing　fault　without　any　pretreatment　and　direct　spectrum　analysis　of　the　vibration　signal　of　fault　bearing.　Through　the　principle　of　resonance　demodulation,　the　resonance　demodulation　method　can　obtain　higher　signal-to-noise　ratio.　This　article　is　taking　the　bearing　field　measured　data　as　the　research　object.　It　proves　that　the　demodulated　resonance　technique　can　improve　the　signal-to-noise　ratio　and　extract　the　fault　information　more　easily.　The　accuracy　of　fault　diagnosis　has　been　improved　greatly　as　well.　©　2013　IEEE.