Pastoral farming systems have always adapted to the seasonal availability of forage resources by moving flocks. However, the role of flock mobility as a possible mitigating strategy in response to climate change has not been clearly documented. Thus, we investigated: (1) the major methodological challenges linked to the diversity of forage resources used and corresponding methane enteric emissions; (2) the functioning of farming systems based on different combinations of natural resources and their GHG emissions; (3) the impacts of grazing practices on soil and biomass carbon flows. We developed an approach combining two existing models (OSTRAL and CASA) enhanced and adapted. This approach was applied to three French Mediterranean sheep farming systems with different degrees of flock mobility: sedentary (SED), single transhumance (ST) and double transhumance (DT). Results produced by OSTRAL whole farm model showed that, considering GHG emissions, DT is the most efficient system with 27,6 kg CO2Eq/kc carcass weight vs 35.1 and 30.9 respectively for ST and SED. Indeed, low inputs and high carcass weight compensate the lower ewe productivity. Non-renewable energy consumption (LCA approach) is very low for DT as there is nearly no inputs: 31 MJ/kg CW vs 82(ST) and 76 (SED). The CASA model allows us to simulate long term carbon balance scenarios according to land cover dynamics in natural environments relatively to grazing intensities, and so to evaluate contribution of livestock to carbon regulations. Combining both models shows that changing the way of using natural grazed areas, and so their capacity of carbon sequestration, could drastically modify hierarchies between these systems concerning mitigation potential. This underlines the potential interest of coupling tools to enlarge the range of analysis.