Maximum Power Point Trackers (MPPT) are widely used to track in real-time the optimal power output of dynamic systems. These systems are sometimes comprised of multiple units which are similar, but not necessarily identical in terms of power curve and dynamics. A good example of such a system would be a photovoltaic (PV) array, which consists of multiple PV cells. Hence, it can be more profitable to operate each unit to its own optimal operating point instead of operating the whole system to a common optimal operating point. This paper proposes to use Particle Swarm Optimization (PSO) as an MPPT where each particle is assigned to a unit of a system. Although the method is validated both through simulations of a PV model and experimentations using a test bench of PV cells, it can be applied to many different dynamic systems comprising multiple units. The method proved to improve the convergence rate of the system and its total power production.