Data envelopment analysis (DEA) is a very effective management tool in assessing the performance of a set of decision making units (DMUs). In the efficiency evaluation using classic single stage DEA models, the internal processes of the DMUs are often neglected. In most real-world problems, it may be more realistic to evaluate the efficiency evaluation in two-stage production systems. In some cases: however, the decision-maker must need to identify the most efficient single unit. Numerous methods have been introduced to find the most efficient unit in single stage systems whereas no methods have been proposed for this aim in two stage production systems. Therefore, a new model based on mixed-integer programming was proposed to determine the most efficient DMU in two-stage systems and sub-stages in this study. The most important innovation of the suggested approach is that the most efficient DMUs of both stages can be found separately using only one model. Numerical examples for real world problems and a simulation study were provided for the validity of the proposed model.