A　Naive　Bayes　algorithm　based　on　the　rough　set　approach　is　proposed　in　this　manuscriptand　successfully　applied　in　the　hospital　medical　caseclassification.　The　core　of　the　algorithm　is　to　filter　the　redundant　data　and　parameterize　the　key　data　in　the　medical　cases　through　the　approaches　ofattribute　reduction　and　system　attribute　significance　in　the　rough　set　theory,　and　calculate　the　weights　in　order　to　achieve　more　satisfactory　accuracyof　medical　case　classification　through　the　Naive　Bayes　algorithm.　The　specific　steps　are:　firstly　removing　the　redundant　phrases　from　the　medical　cases　by　attribute　reduction　of　the　rough　set,　then　calculating　the　weight　of　each　attribute　based　onsystem　attribute　significance,　and　finally　constructing　the　Naive　Bayes　algorithm　model　based　on　the　weighted　system　attributes.　Experiments　have　shown　that　the　above　rough　set　approach-based　Naive　Bayes　algorithm　model　is　a　better　classification　model　that　achieves　higher　classification　accuracy　than　other　Naive　Bayes　algorithm　models.