There　are　many　sensors　used　to　monitor　the　quality　of　the　effluent　during　the　wastewater　treatment　process.　So　the　normal　monitoring　of　the　sensor　is　critical　to　wastewater　treatment.　In　this　article,　the　proposed　sensor　fault　diagnosis　method　is　based　on　fault　diagnosis　of　interval　prediction　which　using　RBF　neural　network　with　set　membership　estimation.　After　some　input　and　output　data　of　the　WWTP　are　obtain,　an　interval　containing　the　actual　output　of　the　system　without　a　fault　can　be　easily　predicted.　If　the　sensor　measured　is　out　of　the　predicted　interval,　it　can　be　determined　that　a　fault　has　occurred.　This　paper　also　establishes　two　independent　interval　diagnosis　models　to　further　make　sure　whether　the　senor　is　faulty　or　the　system　is　faulty.　The　results　demonstrate　that　the　proposed　sensor　fault　diagnosis　method　is　effective　and　useful.　©　2019　IEEE.