In recent years, various models of polyurethane foam have been modeled using CAE analysis. Material properties are important for accurate modeling of polyurethane flow. However, it is difficult to measure accurate polyurethane material properties. Therefore, in CAE analysis, materialspecific parameters are incorporated in the density evaluation equation to correlate with the experimental results. However, since there is a little detail on how the material-specific parameters are determined, it is necessary to construct a method for determining the optimum parameters. In this study, we propose the genetic algorithm optimization method (real-coded GA) that can handle complex nonlinearities. In order to verify the validity of real-coded GA, the time-density evaluation empirical equation proposed by Kono et al. (2005) was used to find the optimal parameters for the material-specific parameters. It was possible to obtain optimum material-specific parameters that fit well with the experiment results. Furthermore, a three-dimensional fluid flow analysis was performed by incorporating Kono's empirical formula. The filling flow behavior and the peak height of the foam obtained in the experiment agreed with the CAE analysis results. It was shown that accurate CAE analysis is possible by this method.