The modeling accuracy influences the performance of model-based control. However, a conventional method does not accurately identify a plant model, and includes modeling errors generated through approximations. In this study, the authors propose a process of modeling for application to a control system design. The proposed method consists of three parts. First, an impulse response estimation method using a time-stretched pulse is applied to elucidate the frequency characteristics of the plant. Next, a complex cepstrum analysis conducted to reduce quantization errors of the estimation result is described. Finally, the estimated frequency characteristics are modeled using BiQuad filters. The validity of the proposed method is confirmed through a numerical simulation of an integral­type servo control. The results show that the proposed method suppresses the transient vibrations of the responses and their external disturbance responses.