To　effectively　identify　the　diamond　sharpness　ratio　during　wheel　dressing　process,　considering　the　non-stationary　specialties　of　acoustic　emission　signals,　the　diamond　blunt　feature　extraction　method　based　on　bispectrum　analysis　was　studied.　The　on-line　wheel　dressing　monitoring　experiments　were　made　to　investigate　the　bispectrum　feature　of　acoustic　emission　signals　when　the　diamond　sharpness　ratio　was　sharp,　middle　blunt　and　blunt,　and　a　diamond　blunt　feature　extraction　method　based　on　main　diagonal　slice　of　normalized　bispectrum　magnitude　of　acoustic　emission　signals　was　proposed.　The　conclusion　is　reached　that　the　diamond　sharpness　ratio　can　be　identified　effectively　by　the　bispectrum　analysis　of　acoustic　emission　signals　during　wheel　dressing　process.