Proton exchange membrane (PEM) fuel cells still suffer from the challenge of limited durability, hindering their widespread commercialization. To overcome this limitation, resorting to Multi-stack Fuel Cell (MFC) systems instead of single fuel cells is a promising solution. Indeed, by optimally distributing the power demand among the different stacks while taking into account their state of health, an efficient Energy Management Strategy (EMS) can be achieved. Here a new multi-stack configuration, based on an oversized multi-stack system is explored. The problem addressed in this paper is to develop a methodology that manages the operation of an oversized three-stack system where only two of them operate simultaneously. The first stage is to predict the deterioration rate of each stack according to the load allocation, and link the deterioration rate of each stack with the load dynamics. To that end, several stochastic deterioration models, from the classical Gamma process model to more complex models with random effects have been developed and tailored to the fuel cell specificities. Then, an event-based decision-making strategy has been established, that determines the load allocations among the operating stacks. This strategy is based on the minimization of the deterioration phenomena due to both the load amplitude and the load variations. Finally, this strategy is extended to the three-stack oversized system by adding the decision to start or stop a stack. These strategies have been validated under random dynamic load profiles, and Monte Carlo simulation results verify the efficiency of the proposed strategies through improved system lifetime.