In practical application, a nominal model is often used to approximate the design of industrial system. This approximation could make the traditional design method less effective due to the existence of model uncertainty. In this paper, a novel robust design approach is proposed to design the robustness of the dynamic system under model uncertainty. The key idea of this proposed method is that it integrates the advantages of both the model-based dynamic robust design and the data-based uncertainty compensation. A simulation example is conducted to demonstrate the effectiveness of the proposed robust design method.