The assumption of classical Data Envelopment Analysis (DEA) models is based on complete homogeneity of Decision Making Units (DMUs). The objective of this paper is to propose a method of determining relative efficiency of slightly non-homogeneous DMUs by using DEA. First missing values are inserted by series mean. Then relative weights of DMUs are measured by Analytic Hierarchy Process (AHP) and finally relative efficiency of DMUs is computed by chance-constrained DEA. A case study demonstrates the application of the proposed method.