In this work an uncertainty quantification procedure and a robust design method are applied to the thermal design of a skin panel and a wing box section. The uncertainty quantification method selected is based on Monte Carlo simulations and a set of common design parameters have been considered and formulated with random values. Numerical results show that the most significant randomness comes from thermal loads and also that the amount of uncertainty decreases from input to output. On the other hand, the Taguchis method of robust design has been applied to obtain the most appropriate values of a set of control factors in several design situations. In this case, the results show and improvement in the response of the components, maintaining a low dependence on input value variations and producing designs that increase the quality level.