This paper deals with multiple constraints for dimension and stacking-sequence optimization of a blade-stiffened composite panel. In a previous study, a multiple objective genetic algorithm using a Kriging response surface with a buckling load constraint was the target. The present study focuses on dimension and stacking-sequence optimization with both a buckling load constraint and a fracture constraint. Multiple constraints complicate the process of selecting sampling analyses to improve the Kriging response surface. The proposed method resolves this problem using the most-critical-constraint approach. The new approach is applied to a blade stiffened composite panel and the approach is shown to be efficient.