Target　setting　is　an　important　reason　that　data　envelopment　analysis　(DEA)　is　widely　applied　to　analyze　efficiency.　However,　the　efficiency　overestimation　of　traditional　DEA　may　affect　the　rationality　of　target　setting,　and　even　may　make　this　method　invalid　in　target　setting.　This　paper　innovatively　proposes　the　definition　of　cross　efficient,　and　develops　two　new　target　setting　approaches　for　individual　decision-making　unit　(DMU)　and　global　optimization　to　improve　the　cross　efficiency　of　DMUs　in　different　decision-making　situations.　Then　the　influence　of　decision　maker＇s　preference　on　the　improvement　direction　is　discussed,　and　then　these　approaches　are　further　extended.　This　paper　solves　the　limitations　of　traditional　DEA　in　target　setting,　and　provides　different　directions　to　improve　cross　efficiency　for　meeting　decision　makers’　needs.　Finally,　two　examples　are　provided　to　illustrate　the　effectiveness　of　these　approaches.　©　2019　Elsevier　Ltd