Computer simulation plays an important role in modern technology development. In particular, it is used for exploring new methodology, examining developed technologies and so forth. Design of experiments can benefit from computer simulation because it is a series of techniques for performing experiments efficiently and the above-mentioned exploration and examination requires experimental results of simulation. This paper discusses the application of design of experiments to computer simulation as a grammar of design of experiments in computer simulation. Specifically, the grammar described in this paper includes techniques for validating a tentative model of response and factors, screening many factors to identify active factors and approximating the response function by active factors. Traditional approaches to design of experiments are partially useful for these areas. However, some of them cannot be solved by the traditional approaches because of the complexity resulting from computer simulation. Therefore, this paper also adopts a new approach to solve these problems.