In　consideration　of　greyness　in　data　of　product　quality　management　for　tobacco,　the　notion　of　data　resource　is　proposed　in　this　paper.　The　available　data　are　purified　from　the　disorderly　and　unsystematic　data　of　quality　managing　via　the　notion　of　data　resource.　Those　purified　data　are　used　to　GM(1,1)　modeling　in　order　to　full　the　blank　of　raw　data,　then　analyzing　and　predicting　the　qualified　rate　of　output　temperature　of　cut　tobacco.