It　has　been　proved　that　higher　total　nitrogen　removal　efficiency　could　be　obtained　in　step　feeding　activated　sludge　process　by　adjusting　the　inflow　rate　distribution　and　the　volume　of　each　stage.　But　the　optimization　of　the　step　feeding　process　would　be　complicated　owing　to　adjustment　of　inflow　rate　and　volume.　To　develop　the　auto-control　and　operational　decision　support　system　is　an　urgent　and　necessary　work　for　well　application　of　step　feeding　process.　Although　biological　wastewater　treatment,　especially　the　biological　nitrogen　removal　process,　has　the　characteristics　of　complex　nonlinear,　it　provides　widespread　opportunities　for　artificial　intelligent　control.　In　this　paper　a　fuzzy　control　model　of　anoxic/oxic　step　feeding　activated　sludge　process　was　developed　under　the　condition　of　sufficient　and　insufficient　influent　carbon　source.　The　detailed　information　on　development　of　fuzzy　model　was　addressed　based　on　collecting　and　analyzing　previous　experimental　data.　The　influent　C/N　ratio　was　selected　as　input　variable　and　inflow　rate　and　volume　distribution　in　each　stage　were　selected　as　output　variables.　Moreover,　the　calibration　and　application　approaches　of　control　model　were　also　studied　through　pilot-scale　experiments　(1000　L/d)　with　domestic　sewage.　The　highlight　is　that　the　method　and　approach　of　model　calibration　are　well　discussed　so　that　it　provides　good　opportunities　on　practical　application.　During　three　months　experimental　period　enhanced　total　nitrogen　removal　efficiency　of　more　than　95.1%　was　reached　with　the　influent　C/N　ratio　of　8.16-12.3,　average　of　9.78.　(C)　2009　Published　by　Elsevier　B.V.