According to European Union strategies, hydrogen technologies have a significant potential for the decarbonization of the automotive sector. Fuel Cells are considered a highly sustainable alternative to internal combustion engines for hybrid powertrain solutions. Since experimental tests on real prototypes are extremely costly in terms of time and resources, they represent a limit to the development rapidity of such complex vehicles. Consequently, simulation models are gaining further importance for their intrinsic time- and cost-saving characteristics, while their predictive capability is crucial. Accordingly, the development of the so-called “digital twins” able to accurately represent the real-time digital counterpart of a physical system has become an important research issue. As they allow to account for the impact of parameters such as driving cycles, they may lead the design of the system core components as well as of the overall Balance of Plat, also taking into account the role of auxiliaries, that are often neglected. Accounting for these effects can significantly influence the power splitting between fuel cell and batteries, ultimately affecting the vehicle performance. In the present work the impact of driving cycles on the design of advanced control strategies is addressed. To this aim, a digital twin of a light-duty commercial vehicle is tested under different operating conditions, representing an urban and extra-urban driving cycle. Two optimizations on the control have then been carried out by means of genetic algorithm to minimize the hydrogen consumption. Results highlight that adapting the power splitting control logic to the specific operating conditions can lead to saving up to 4 % on the overall vehicle consumption, due to a different way of managing the power split and the battery state of charge. Moreover, the two sets of optimized parameters have been tested on a random driving cycle in order to highlight the differences in behavior.