Abstract: Based on the robust optimization approaches,several robust data envelopment analysis (RDEA) methods are proposed in this paper so as to deal with the negative impact of uncertain information to the evaluation results of traditional DEA method.The RDEA models are implemented in a numerical example and the efficiencies and rankings of these models are compared.The evaluation and sensitivity analysis results indicate that the proposed RDEA models can be immune against inputing and outputing data uncertainties and provid more reliable and stable efficiency evaluation information.