The rapid growth of complexity of communication networks - including grids, cloud systems and services - increases the need for the use and performance prediction of parallel and distributed discrete event simulation. The paper aims at the decreasing the overall cost of simulation by improving the performance prediction of simulation particularly for the phases with uncertain and vague information. Involving rough sets-based methods, the paper introduces how the efficiency of the well known coupling factor method of the performance prediction of parallel and distributed simulation can be improved. According to the presented results, the prediction can merely be based on the maintainable lookahead feature of the designed simulation model. Using the rough set analysis, the paper also describes how the number of simulation experiments necessary for prediction can be limited without the decrease of accuracy of prediction.