On　the　basis　of　analyzing　the　classical　methods　of　sludge　process　modeling,　the　paper　put　forward　a　new　method　about　activated　sludge　process　by　Neural　Networks.　Firstly,　the　paper　utilized　principal　component　analysis　method　to　realize　reduce　the　dimension　of　the　input　vectors　and　orthogonalize　the　components　of　the　input　vectors.　Then　built　activated　sludge　process　system　by　BP　and　RBF　Artificial　neural　networks,　the　applicability　of　the　two　neural　network　models　were　analyzed　to　sludge　process.　The　experiment　result　shows　that:(1)　These　neural　networks　may　reflect　real　conditions　correctly　and　have　strong　self-adaptation;(2)　The　RBF　neural　network　model　has　better　convergence　ability　and　impending　speed　than　the　BP　neural　network　model.