The increasing level of complexity of energy systems drives researchers to focus their studies on energy optimization, by using modelling and simulation methods capable to represent the real system behavior. In this study, a functional energetic modeling method is used to design a control architecture for energy flow management, which relies on local control loops, a decision manager (DM) and basic equations. When the functional level of representation is used to model a complex system, the evaluation of model accuracy (from an energetic point of view) and the validation of energy management algorithms are eased by fast simulations due to low model complexity. While the functional model allows a first-stage validation of energy distribution within the system, the energy management algorithms need to be tested using a more accurate model, which is the multi-physical model of the system. The multi-physical model has its own local controllers and a global resource manager (GRM) to handle the power split between different components. The second-stage validation can be completed by adapting the functional model in order to design the high-level controller, the GRM, at multi-physical level. To develop the control architecture of the multi-physical model based on the functional model, two steps are required: i) adjust the parametrization of functional elements and ii) propose a method to interconnect the models at both levels of representation (functional/ multi-physical level). Thus, an example of a hybrid electric vehicle (HEV) is considered for functional elements modelling and parametrization. In addition, the GRM design is presented and simulation results of the HEV system at multi-physical level are illustrated to validate the system architecture and component sizing, and to evaluate the fuel consumption compared with HEV design specifications.