Abstract : Hybrid vehicles are usually propelled by one or two electric motor and an internal combustion engine. The first part of this study presents some classical mechanical arrangements and describes the prototype built at the LAMIH during a previous project. Algorithms, called control strategy, are used to choose at each sampling time the power split between the engine and the motor. Some of them rely on empiric rules and uses fuzzy controllers and other deals with more detailed energy flow analysis. Then the choice of the power split is written into the form of an optimization under constraint problem. Several algorithms based on the optimal control theory are proposed. The aim is to compute, in simulation, the control sequence that ensures the lowest fuel consumption. A new control strategy is derived from the global optimization algorithm and a second one is proposed, based on the analysis of the global optimization results. Several criterions are proposed in order to characterize speed cycles, to evaluate control strategies. A comparison of several control strategies shows that one of the proposed control strategies provides best results, closed to those of the global optimization algorithm. At last, the use of the global optimization algorithm for the choice of the mechanical arrangement and sizing the engine and the motor is explained in the last section.