On-going oil stock depletion and growing environmental concerns lead automakers to develop more efficient powertrains. Today the most promising way forward consists in research on hybrid systems. The present study uses TRIZ methodology to help identify the best hybridization architecture and powertrain design, in order to reduce CO2 emissions. This optimization problem is constrained with cost, mass and complexity targets. It is applied to B segment hybrid electric vehicle (HEV) with no connection to the grid to recharge the battery. A specific procedure based on TRIZ principles has been established and rolled out after a first mathematical analysis of different current hybridization architectures (series, parallel and combined). The adapted TRIZ methodology is mainly built on the principles of Ideal Final Result and on networks of contradictions. It is used to compare the performances of the different architectures, identify the main technological barriers for more CO2 savings, and propose a roadmap to overcome them.