The design of experiment (DoE) technique and one of its aspects, the dual response surface (DRS) method, are employed to analyze the results of an ageing test performed during 1000 hours on a PEM fuel cell stack made up of three cells. The response surfaces are plotted thanks to load current - fuel cell voltage curves recorded at regular time-spaced intervals and for various air utilization rates. The use of the DRS approach in the determination of a robust fuel cell control is here appropriate to select the most convenient operating conditions for the cells (load current, air stoichiometry rate), leading to a trade-off between some high performances and a low variability versus ageing time and cell positions in the stack