Because　the　technology　of　classifiers　ensemble　can　not　only　improve　the　performance　of　weak　classifiers　effectively　but　also　can　be　added　conveniently　into　existing　text　classification　framework,　it　has　been　widely　used　in　text　classification　and　filtering　field.　After　analysis　of　current　technologies,　the　general　framework　of　spam　filtering　was　given.　In　addition,　the　technologies　of　spam　filtering　based　on　classifiers　ensemble.　Then　an　approach　of　classifiers　ensemble　based　on　corpus　partition　was　proposed.　Experimental　results　show　that　the　approach　can　improve　the　accuracy　of　text　classification.　Besides,　the　resulting　technology　has　been　successfully　implemented　in　the　spam　filtering　system.