Abstract: Traditional fail-safe optimal design was always performedbased on a prior assumed failure mode. However, the assumed failure mode maynot necessarily be the most dangerous one, and if this is the case, then thereliability of the obtained optimal solution cannot be guaranteed. Tocircumvent this difficulty, in the present paper, fail-safe optimal designis carried out in a robust optimization framewok. The problem is formulatedin a Bi-level form. The upper level program aims at finding the optimaldesign variables while the lower level program is solved to find the worstcase failure mode. In order to find the global optimal solution of the lowerlevel program, a mixed 0-1 programming is proposed and solved with thebranch-and-bound algorithm. The effectiveness of the proposed approach forfail-safe design and optimization are illustrated by some numericalexamples.