Aiming　at　the　difficulty　of　separating　the　fault　feature　from　compound　rolling　bearing　fault　signal,　a　new　fault　diagnosis　method　is　proposed　based　on　dual-tree　complex　wavelet　transform　(DT-CWT)　and　independent　component　analysis　(ICA).　Firstly,　DT-CWT　is　used　to　decompose　the　non-stationary　fault　vibration　signal　into　several　components　with　different　frequency　bands.　Because　frequency　aliasing　exists　in　the　components,　this　problem　disturbs　the　feature　extraction　of　the　fault　signal.　Then,　ICA　is　introduced　to　perform　blind　source　separation　on　the　mixed　signal　consisting　of　various　components　to　eliminate　the　frequency　aliasing　as　far　as　possible.　Finally,　Hilbert　envelope　decomposition　is　performed　on　the　independent　signal　components　separated　from　the　mixed　signal.　Thus　the　compound　fault　feature　information　can　be　separated,　and　the　fault　identification　is　achieved.　The　experiment　results　show　that　the　proposed　method　can　effectively　separate　and　extract　the　feature　information　of　the　compound　rolling　bearing　faults,　which　verifies　the　feasibility　and　effectiveness　of　the　proposed　method.