In　terms　of　the　problem　of　ranking　decision　making　unit　(DMU)　in　data　envelopment　analysis,　an　evaluation　approach　based　on　cross　ranking　matrix　of　DMUs　is　established.　On　the　basis　of　neutral　cross　efficiency,　this　matrix　can　be　obtained　from　relative　ordinal　in　the　corresponding　cross　efficiency.　Then,　the　sum　of　efficiencies　for　each　DMU　on　different　sequences　can　be　gained　by　aggregating　the　self-　and　peer-evaluation　values.　Next,　a　linear　programming　model　is　established,　and　the　DMUs　are　ranked　according　to　the　results　of　this　model.　The　proposed　method　is　applied　to　the　efficiency　evaluation　of　several　experimental　demonstration　centers　in　Fuzhou　University　town.　The　result　shows　that　our　model　works　better　than　CCR　and　NCE　in　distinguishing　the　performances　of　DMUs.　©　2018,　Zhejiang　University　Press.　All　right　reserved.