In　a　category　of　production　processes,　it　is　difficult　to　realize　the　product　quality　indices-oriented　optimizing　control　due　to　their　characteristics　of　multi-variable　and　un-measurable　product　quality.　To　deal　with　such　a　problem,　an　intelligent　evaluation　system　for　the　set　points　in　the　control　loops　is　proposed　on　the　basis　of　the　methodology　of　case-based　reasoning　(CBR).　Consisting　of　a　case-based　reasoning　model,　an　off-fine　index　measuring　process,　and　a　self-tuning　model,　this　system　achieves　real-time　evaluation　and　closed-loop　optimization　for　the　set　points.　The　proposed　evaluation　system　is　applied　to　a　practical　industry　process,　where　it　can　exactly　evaluate　the　rationality　of　the　set　points　under　the　scenarios　with　frequently　varied　information.　Since　the　product　quality　can　be　predicted　with　required　accuracy,　the　product　indices-oriented　optimization　is　fulfilled,　bringing　in　remarkable　profits.