The　paper　put　forward　a　method　based　on　GM(1,1)-SVM　for　rolling　bearing　fault　prediction　and　diagnosis.　Firstly,　time　and　frequency　domain　feature　values　of　vibration　signal　of　rolling　bearing　under　all　kinds　of　fault　and　normal　condition　were　extracted.　Then　the　important　characteristic　parameters　were　collected　to　build　the　predict　model.　Lastly,　fault　and　normal　condition　eigenvalue　was　used　to　train　binary　tree　support　vector　machine　and　to　construct　the　decision　tree　to　classify　the　fault　type.　Thus　the　bearing　fault　diagnosis　and　the　fault　prediction　through　the　predicted　values　and　the　support　vector　machine　(SVM)　were　achieved.　©　2015,　Beijing　University　of　Technology.　All　right　reserved.