Due　to　the　complex　dynamic　behavior　in　wastewater　treatment　process,　online　measurement　of　ammonia　nitrogen　value　is　very　difficult.　In　this　paper,　a　case-based　reasoning　(CBR)　prediction　model　based　on　a　feedforward　neural　network　(FNN)　is　introduced　to　predict　the　effluent　ammonia　nitrogen　value.　First,　easily　measured　feature　variables　which　have　great　effect　on　effluent　ammonia　nitrogen　value　were　selected.　Next,　the　prediction　model　was　established,　and　attribute　weights　in　case　retrieval　were　determined　by　the　connection　weights　of　a　trained　FNN.　Finally,　based　on　the　data　in　a　real　wastewater　treatment　process,　simulation　experiments　were　carried　out.　The　results　show　that　the　prediction　model　using　FNN-based　CBR　is　effective　and　has　better　prediction　accuracy　than　some　other　methods.　©　2018　Technical　Committee　on　Control　Theory,　Chinese　Association　of　Automation.