Municipal　solid　waste　incineration　(MSWI)　technology　has　been　widely　used　all　over　the　world.　As　fluctuation　in　the　solid　waste　components,　its　operation　usually　depends　on　the　domain　experts　to　judge　the　combustion　state　in　the　furnace.　The　key　operating　variables,　such　as,　air　volume,　are　manually　set　according　to　expert　experience,　which　has　strong　subjectivity　and　randomness.　To　solve　the　above　problems,　this　paper　builds　an　air　volume　setting　model　of　MSWI　process　based　on　color　moment　features　of　combustion　flame.　First,　single　image　fast　defogging　and　median　filter　algorithms　are　used　to　defog　and　eliminate　random　noise.　Then,　the　color　moment　features　of　the　preprocessed　image　are　extracted　by　multi-scale　sliding　windows.　Thirdly,　the　correlation　between　color　moment　features　and　air　volume　is　analyzed　by　mutual　information　algorithm　(MI)　for　selecting　the　key　features.　At　last,　the　air　volume　setting　model　based　on　radial　basis　function　neural　network　(RBFNN)　is　constructed.　The　effectiveness　of　the　proposed　method　is　verified　by　the　actual　running　data　of　a　MSWI　power　plant　in　Beijing.　©　2021　IEEE