Addressing　the　challenge　of　precisely　controlling　furnace　temperature,　this　study　introduces　a　modeling　strategy　based　on　an　enhanced　hierarchical　fused　fuzzy　deep　neural　network　(HF-FDNN)　model.　Initially,　an　analysis　of　the　control　characteristics　of　furnace　temperature　identifies　key　manipulated　variables　(MVs).　Subsequently,　we　refine　the　HF-FDNN　algorithm　within　the　task-driven　layer　to　develop　a　model　for　furnace　temperature　control.　Finally,　we　validate　the　effectiveness　of　proposed　method　through　experimental　results　derived　from　real　municipal　solid　waste　incineration　(MSWI)　process　data.　©　2024　IEEE.