In general, in probabilistic seismic hazard analysis, the ground motion prediction equation (GMPE) is suitable for evaluating ground motions from fault sources with a single dipping angle. However, for the complex fault rupture cases of listric faults that have two or more dipping angles, it is difficult to use the GMPE for seismic hazard analysis because of the nature of its functional form. The overall characteristics of ground motion at a specific site of interest need to be determined through multiple earthquake scenarios via ground motion simulations. Therefore, it is necessary to compare ground motions from GMPE prediction and simulated data in order to set the GMPE for use in seismic hazard analysis. In this study, we employ a ground-motion simulation to evaluate GMPE input parameters for listric normal fault cases. We show that quantitatively comparing ground motions from GMPE prediction and simulated data is a feasible method for defining input parameters for the GMPE in seismic hazard analysis.