To　improve　the　control　performance　of　the　setting　model　with　multi-objective　evaluation　and　case-based　reasoning　(MOE　&　CBR)　for　shaft　furnace　roasting　process,　we　make　use　of　the　water-filling　based　weight　allocation　(WFA)　to　allocate　weights　for　process　variables　and　employ　the　group　decision-marking　revision　(GDMR)　to　develop　a　new　intelligent　setting　method.　First,　a　Lagrange　function　is　constructed　to　optimize　the　allocation　of　the　weights　of　the　process　variables　via　WFA.　Subsequently,　the　suggested　solutions　of　set-points　are　obtained　through　case　retrieval　and　case　reuse.　These　suggested　solutions　are　used　to　evaluate　the　production　performance　indices　based　on　the　multi-objective　evaluation　(MOE)　model.　Those　unreasonable　set-points　from　MOE　model　are　revised　by　GDMR.　The　proposed　method　has　been　applied　to　the　shaft　furnace　roasting　process.　The　application　results　indicate　that　the　proposed　method　is　superior　to　other　methods　and　it　can　significantly　improve　the　control　performance　of　MOE　&　CBR　model.　©,　2015,　South　China　University　of　Technology.　All　right　reserved.