E-government　is　full　of　all　kinds　of　information　about　government　affair.　Faced　with　this　overload　of　information,　the　public　will　make　wrong　decisions　or　begin　to　feel　uncertain.　To　reduce　information　overload　on　e-government,　the　idea　of　personalization　services　has　emerged.　In　this　paper,　we　put　forward　a　public-oriented　model　of　personalized　information　services　based　on　data　mining　in　e-government,　discuss　the　key　techniques　adopted　in　the　model,　and　give　the　implement　processes　for　them.