Rolling　bearing　is　one　of　the　main　parts　of　rotating　machinery,　but　the　complex　and　changeable　working　environment　causes　frequent　failure　and　many　kinds　of　composite　fault.　In　order　to　solve　this　problem,　a　composite　fault　diagnosis　method　of　rolling　bearing　based　on　improved　maximum　correlated　kurtosis　deconvolution　(MCKD)　and　teager　energy　operator　is　proposed.　In　this　method,　particle　swarm　optimization　(PSO)　is　used　to　optimize　the　MCKD　parameters　(L　and　M)　of　different　types　of　faults,　set　up　the　deconvolution　period　corresponding　to　the　fault　type,　calculate　the　MCKD　algorithm　with　the　maximum　correlation　kurtosis,　and　improve　the　filter　coefficients.　The　improved　MCKD　algorithm　reduces　the　noise　interference　to　a　great　extent,　and　then　use　the　teager　energy　operator　to　detect　the　transient　impact　of　the　signal,　and　analyze　the　teager　energy　spectrum　to　realize　the　composite　fault　diagnosis.　Finally,　the　method　is　validated　by　using　the　bearing　data　of　Case　Western　Reserve　University　and　the　bearing　fault　simulator,　and　the　results　show　that　it　can　effectively　extract　fault　feature　information　from　single　and　composite　fault　of　rolling　bearing　and　identify　the　fault　type　accurately.　©　2019,　Editorial　Office　of　Journal　of　Dalian　University　of　Technology.　All　right　reserved.