A structure, whose characteristics are less modified by any changes of external forces, boundary conditions and productive errors, has advantages for its practical use. This character is called robustness. In the consecutive study (Yamakawa and Miyashita, 1991), the authors studied the structural design with robustness and proposed a new structural concept called “Robust Structure.” Robust structural design is more useful when the structures are designed in the upper flow of the design process than in the lower flow. Robust structures can be designed by minimizing the sensitivities of the reference function relative to the variable parameters. However, the sensitivities provide only the local information even when their values are evaluated. Therefore, it is not clarified whether robust structure designed by the sensitivities has wide robustness or not. To conquer this problem, this paper shows robust structural design method using the fuzzy sets. In this method, the variable parameters are shown with fuzzy sets and then obtain the wide robustness. It is also pointed out that robust structural design using sensitivities can design wide robust structure in certain cases. Because robust structural design using fuzzy sets needs a lot of numerical iterations, it is important to figure out the effective design method of wide robust structures. In this paper, an idea is proposed how to design wide robust structures. The numerical examples designed by two different methods using sensitivities and fuzzy sets are shown. Further more, the proposed idea is discussed with these examples.