In　recent　years,　the　application　of　Web　technology　is　more　and　more　mature,　and　all　manner　of　Websites　and　search　engines　are　distributed　on　the　Internet　resulting　in　a　huge　source　of　information.　Thus,　if　the　Web　information　is　not　well　organized,　it　is　hard　to　find　the　information　what　a　user　actually　need　or　interest　in.　Automatic　classification　of　Web　pages　is　an　efficient　method　in　Web　Content　Mining　which　can　be　of　great　value　in　the　management　of　Web　directories.　Based　on　the　analysis　done,　The　Cerebellar　Model　Articulation　Controller　(CMAC)　is　an　excellent　classification　technique,　but　when　it　is　applied　to　deal　with　high-dimensional　dataset　such　as　the　data　on　the　Internet,　the　memory　required　increase　intensively.　This　paper　presents　an　improved　CMAC　model　used　in　content　based　Web　page　classification　which　requires　less　memory.　The　experimental　results　show　that　the　proposed　model　is　highly　effective　in　Web　page　classification.