An simulation has become a major tool in studying discrete manufacturing systems, experimental design issues have started to receive much attention. Although the importance of proper experimentation is often emphasized in the literature on discrete event simulation, it is neglected in most practical simulation studies. The main reason for this neglect is that the design and analysis of a simulation experiment require expertise in experimental design methodology as well as familiarity with traditional statistical output analysis methods. This article presents an attempt to structure the process of planning and designing simulation experiments, and proposes a knowledge-based system, namely Design Of Experiments for Simulation (DOES), that assists the inexperienced analyst in this process. The proposed system takes the analyst through a planning process based on the simulation objectives.