Data envelopment analysis (DEA) can be used for supplier selection problem due to its multiple criteria nature. In suppliers' evaluation, there might be some factors, which are beyond the control of their management, that are needed to be modelled in an appropriate way. Also, there are some situations in which some factors are undesirable and they are favourable to be decreased. The aim of this paper is to propose a model for evaluation of suppliers' performance in the presence of both undesirable and non-discretionary outputs. This model can rank efficient suppliers by a super-efficiency DEA model. A numerical example has sought to demonstrate that the proposed model is actually applicable.Copyright 2014 Inderscience Enterprises Ltd.