As　currently　widely　used　municipal　solid　waste　resource　treatment　method,　the　municipal　solid　waste　incineration　(MSWI)　process　emit　dioxins　(DXN)　compound　s　with　high　toxicity　and　persistent　pollution　characteristics.　It　is　one　of　the　main　reasons　that　cause　incineration　power　plants　to　have　a　＂Not　in　my　back　yard＂.　At　present,　the　long-period,　high-cost　offline　detection　method　used　in　industrial　sites　cannot　achieve　real-time　monitoring　of　DXN　emission　concentration.　Moreover,　the　number　of　samples　used　to　build　a　DXN　emission　concent　ration　prediction　model　is　extremely　scarce.　Aim　at　the　above　problems,　a　method　for　predicting　DXN　emission　concentration　in　MSWI　process　based　on　PSO　and　equi　spaced　interpolation　is　proposed.　At　first,　the　domain　of　original　small　sample　input　and　output　is　expanded　based　on　the　improved　mega-trend-diffusion(MTD)　technology.　Then,　the　equal　interval　interpolation　method　is　used　to　generate　the　virtual　sample　inputs,　which　are　used　to　obtain　the　virtual　sample　output　s　by　combining　the　mapping　model.　The　above　results　are　combined　with　the　expansion　space　to　delete　the　bad　virtual　samples.　Thirdly,　the　PSO　algorithm　is　used　to　optimize　and　select　the　reduced　virtual　sample　set.　Finally,　a　mixed　sample　set　composed　of　the　optimized　selected　virtual　sample　set　and　the　original　small　sample　set　is　used　to　construct　a　DXN　emission　concentration　prediction　model.　The　effectiveness　of　the　proposed　method　is　verified　with　the　actual　detection　data　of　DXN　for　many　years　in　a　MSWI　plant.