An adaptative energy management strategy for series hybrid electric vehicles based on optimized maps and the SUMO (Simulation of Urban MObility) predictor is presented here. The first step of the investigation is the off line optimization of the control strategy parameters (already developed by the authors) over a series of reference mini driving cycles (duration of 60s) obtained from standard driving cycles (UDDS, EUDC, etc) and realistic driving cycles acquired on the ITAN500 HEV. The optimal variables related to each mini driving cycle are stored in maps that are then implemented on the ITAN500 vehicles. When the vehicle moves, a wireless card is used to exchange information with surrounding vehicle and infrastructure. These information are used by a local instance of the SUMO traffic prediction tool (run on board) to predict the driving conditions of the HEV in the future period of time T=60s. The predicted driving cycle is compared with the reference mini driving cycles and the most similar one is found. The optimal control strategy parameters mapped for that reference cycle are then used to select the power-split in the future time window. This process is repeated every T seconds obtaining an adaptative control strategy which do not requires much computational power on board. The proposed approach has been compared numerically with the “no knowledge” approach and the “full knowledge” approach. In the “no knowledge” case, the energy management was optimized for NEDC and then applied to three realistic driving cycles. In the “full knowledge” approach the energy management was optimized for each realistic driving cycle. The “full knowledge” approach allows the best fuel consumption to be obtained but requires the knowledge of the whole vehicle mission while the “no knowledge” method gives poor results since it cannot exploit the potentiality of a PHEV. The proposed approach allows good results to be obtained in terms of fuel consumption thanks to a better usage of the internal combustion engine.