Pastoral farming in sylvo-pastoral systems remains a key activity for the local economy and food security in the Sahel. Livestock has a great impact on the carbon cycle and on greenhouse gas (GHG) fluxes. Over the last decade, experimental and modeling studies have addressed the variability of ecosystem GHG fluxes, their drivers and efforts at scaling. Nonetheless, in these Sahelian semi-arid ecosystems, there is still limited information on the underlying processes as well as temporal and spatial variability of GHG fluxes, preventing their quantification from local-to-regional scales. This lack of information is mostly due to the paucity of experimental data, but also modeling studies required at the regional scale. We used STEP (Sahelian Transpiration Evaporation and Productivity model), an ecosystem process model for Sahelian herbaceous savannas to investigate the underlying processes of carbon dioxide (CO₂) and nitrous oxide (N₂O) emissions in a typical sylvo-pastoral system of the Sahel (Dahra, Senegal). Overall, good agreements have been found between model outputs and validation in-situ data for soil moisture, evapotranspiration, soil mineral and organic N, soil C, and herbaceous mass. Dynamics of the model output of N₂O and CO₂ fluxes clearly show emission pulses after the first rains as generally observed in dry ecosystems. Following calibration and validation at the local scale, the model will be used to assess GHG emissions from different sites across the Sahel. GHG flux maps, validated by in situ measurements, will then be proposed at the scale of the Sahelian region for sylvo-pastoral ecosystems.