Call　center　has　been　paid　more　and　more　attention,　a　method　for　predicting　call　center　service　grade　with　improved　neural　network　algorithm　was　put　forward　according　to　the　call　center　service　quality　management　requirements　in　the　enterprise.　The　optimization　algorithm　Levenberg-Marquardt　was　utilized　to　increase　the　convergence　speed　of　BP　neural　network.　And　overcome　the　shortcomings　of　falling　into　local　minimum　value　easily.　The　data　mining　of　call　center　system　was　realized　offering　a　scientific　and　rational　basis　for　the　decision　analysis　to　improve　the　level　of　call　center　services.　©2010　IEEE.