Accurate　prediction　of　nitrogen　oxide　(NOx)　concentration　is　crucial　in　controlling　NOx　emissions　in　municipal　solid　waste　incineration　(MSWI)　processes.　In　this　study,　a　model　based　on　radial　basis　function　(RBF)　neural　network　and　competitive　swarm　optimization　(CSO)　algorithm　is　proposed　to　predict　the　NOx　concentration　ahead　of　time　with　high　accuracy.　First,　by　considering　the　continuous　change　of　NOx　concentration,　the　embedding　dimension　of　NOx　for　model　inputs　is　determined.　Then,　the　RBF　neural　network　whose　parameters　are　optimized　by　the　CSO　algorithm　is　utilized　to　establish　the　prediction　model,　aiming　to　improve　the　prediction　accuracy.　Finally,　simulation　experiments　are　carried　out　on　the　benchmark　data　and　real　industrial　data.　The　comparison　results　illustrate　that　the　prediction　accuracy　of　the　model　proposed　in　this　study　is　higher　than　basic　RBF　model　and　PSO-RBF　model.　©　2021　IEEE