This paper describes a new model predictive control method for buck DC-DC converters. The objective is to achieve a good trade-off between fast response and low overshoot in the transient response of the output voltage. In the well-known PWM-PI method, such a trade-off is difficult to attain. The proposed model predictive control is based on an analytic algorithm that can be easily implemented online. Further, a PI outer loop is introduced to improve the robustness against load uncertainty. Moreover, the advantage of the proposed model predictive control is verified experimentally. The experimental results show that the proposed method outperforms the PWM-PI method.