Data Envelopment Analysis (DEA) is an effective method for measuring the efficiency of Decision Making Units (DMUs). DMUs have been generated based on production function, cost function and measuring efficiency for evaluation and selection. Prediction of DMUs' Performance (PDMUP) is a problem to select the best among DMUs based on input and output data of the DMUs. PDMUP includes evaluating of ranking and efficiency of DMUs. Improving the accuracy and computation time in measuring the efficiency of PDMUP have been two main challenges. Hence, in this paper, a method is proposed using DEA and a Support Vector Machine (SVM) for predicting DMUs. This method has been compared using datasets in earlier research for large DMUs. The experimental results of comparisons between the PDMUP method and early methods demonstrate that the method can significantly improve the accuracy and reduce the computation time in predicting the efficiency of large DMUs.