Aiming　at　the　problem　that　traditional　methods　with　only　one　minsup　can　not　completely　reflect　different　appearing　frequencies　and　natures　of　different　data　items,　based　on　FP-Tree,　a　new　algorithm　is　proposed　called　MSDMFIA　(Multiple　minimum　Supports　for　Discover　Maximum　Frequent　Item　sets　Algorithm),　The　algorithm　allows　users　to　specify　multiple　minsups　to　reflect　various　items　natures.　Through　mining　the　maximum　frequent　item　sets,　calculating　minsups　of　the　maximum　candidate　frequent　item　sets,　the　association　rules　can　be　discovered.　The　algorithm　resolves　the　bottlenecks　in　traditional　algorithms,　e.g.,　the　rare　item　problem,　the　frequent　generation　of　candidate　item　sets　and　database　scanning.　Experimental　results　show　that　functionality　and　performance　of　the　proposed　algorithm　is　significantly　improved　compared　with　existing　algorithms.