The　data　pretreatment　to　reduce　the　redundancy　of　data　and　eliminate　the　data　imprecise　is　the　bottleneck　in　data　mining.　This　article　uses　the　superiority　of　the　Rough　Set　model　to　eliminate　the　data　imprecise,　reduce　the　redundancy　as　far　as　possible　in　the　data　pretreatment　and　reduces　the　data　capacity　by　the　deduction　of　the　attributes　and　deduction　of　values.　It　has　also　laid　the　foundation　for　the　following　data　mining　algorithm　to　reduce　the　time　of　process.　By　the　example　proof　it　is　one　better　algorithm　according　to　incomplete　and　redundancy　data　processing.