Mining　and　discovering　clusters　from　tremendous　data　is　a　useful　analysis　work　for　many　applications　like　economics,　medicine,　engineering,　etc.　As　a　widely　applied　clustering　method,　Kmeans　has　the　merits　of　fast　running　and　moderate　clustering　quality.　However,　the　traditional　Euclidean　measure　has　its　own　inefficiency.　In　this　paper,　a　new　clustering　method　that　integrates　the　grey　relational　analysis　from　grey　theory　into　Kmeans　algorithm　is　proposed　to　overcome　the　shortcomings　of　traditional　Kmeans.　By　applying　to　the　analysis　of　reginal　competitive　ability　of　regions　in　China,　the　new　algorithm　proved　to　be　an　effective　and　efficient　method.　©　2014　IEEE.