In　uncertain　environment,　the　cross-efficiency　matrix　will　be　expressed　by　various　and　complex　fuzzy　data.　In　order　to　aggregate　the　cross-efficiency　with　different　types　of　fuzzy　data,　the　similarity　theory　of　the　generalized　fuzzy　numbers　is　introduced　to　change　generalize　fuzzy　cross-efficiency　matrix　into　the　cross-evaluation　similarity　matrix.　The　comprehensive　similarity　of　peer　evaluation　of　each　(decision　making　units　DMU)　is　computed　by　ordered　geometric　mean　function,　which　satisfied　the　proposition　of　＇positively　compensatory＇.　Moreover,　based　on　the　comprehensive　similarities　of　peer　evaluation　of　DMUs,　a　fuzzy　additive　consist　preference　matrix　is　constructed　to　yield　aggregation　weights　of　DMUs.　The　final　cross-efficiency　of　DMUs　is　obtained　by　aggregating　the　cross-efficiency　matrix.　The　aggregation　method　in　this　paper　has　wide　adaptability,　and　the　results　of　it　are　highly　consistent　and　stable.　At　last,　this　paper　demonstrates　the　applicability　of　fuzzy　cross　efficiency　aggregation　based　on　the　similarity　by　taking　the　example　of　12　suppliers　from　semiconductor　industry.　©　2020,　Editorial　Board　of　Journal　of　Systems　Engineering　Society　of　China.　All　right　reserved.