The　rough　set　theory　is　a　theory　method.　It　is　about　incomplete　research,　uncertain　expression　of　knowledge　and　data,　knowledge　and　data　discovery.　It　has　been　widely　used　in　artificial　intelligence,　knowledge　and　data　discovery,　pattern　recognition　and　classification,　analysis　and　reasoning　of　imprecise　data　and　find　the　potential　knowledge,　data　mining.　But　the　traditional　rough　set　theory　can　only　deal　with　discrete　attributes　in　the　database.　While　the　vast　majority　of　real　database　contains　both　discrete　attributes　and　continuous　attributes.　Therefore,　we　must　turn　these　continuous　attributes　into　discrete.　Aiming　at　this　problem,　this　paper　implements　a　discretization　method　based　on　genetic　algorithm.　It　is　using　genetic　algorithm　to　obtain　the　target,　while　keeping　the　original　decision　system　under　the　condition　of　the　indiscernibility　relation　and　minimum　break　point　set　as　the　optimization　target.　Experimental　results　show　that　the　discrete　method　using　genetic　algorithm　is　feasible.