针对传统数据包络分析的排序主要是基于决策单元的自评效率,决策单元使用的是对自身最有利的权重而使效率不具有可比性,提出一种新的决策单元排序方法.该方法通过构造有利于整体决策单元效率的权重集合,在交叉效率模型中对权重进行约束,从而有效地减小他评效率与自评效率的差距,使排序结果更合理,也更易被决策单元接受.通过算例说明该方法的有效性. The traditional data envelopment analysis is mainly based on the self-assessment efficiency of decision-making units, decision-making unit is using their own most favorable weight so that the efficiency is not comparable, a new decision-making unit sorting method is proposed by Constructing a set of weights that is conducive to the efficiency of the overall decision making unit and constraining the weights in the cross efficiency model so as to effectively reduce the gap between his evaluation efficiency and self-evaluation efficiency, making the sorting result more reasonable and more acceptable to the decision making unit. An example illustrates the effectiveness of this method.