Efficiency and energy consumption reduction are becoming a key issue in elevation applications. Energy Storage Systems (ESS) can play a significant role on this field, together with their associated Energy Management System (EMS) in order to optimize the overall behavior of the elevator. This paper presents an EMS based on Dynamic Programming (DP) for a stochastic application, introducing a novel representation of these systems (General Energy and Statistical Description) and implementing a cost function based on the stock management theory. It has been implemented and validated experimentally on a real elevator with energy storage capability reducing grid power peaks by 65% and braking resistor energy losses up to 84%.