This paper addresses the issue of optimal sizing reliability applied to a fuel cell/battery hybrid system. This specific problem raises the global problem of strong coupling between hardware and control parameters. To tackle this matter, the proposed methodology uses nested optimization loops. Furthermore, to increase the optimal design relevance, a reliability assessment of the optimal sizing set is introduced. This new paradigm enables showing the early impact of the reliability criteria on design choices regarding energetic performance index. It leads to a smart design methodology permitting to avoid complexity and save computing time. It considerably helps design engineers set up the best hybridization rate and enables practicing tradeoffs, including reliability aspects in the early design stages.