Effective grassland management is crucial for carbon sequestration and livestock productivity. Our study improved the ModVege model by adding spatially explicit grazing simulations to see how differ- ent grazing methods impact vegetation heterogeneity. We looked at four grazing approaches: uniform defoliation, random grazing, palatability-based selection and Optimal Foraging Theory. We tested these across various stocking rates (approximately 0.7–2 livestock units per hectare per year) and man- agement regimes (continuous vs. rotational). Our results show that heterogeneous grazing patterns, such as random defoliation, increase spatial variance and help preserve reproductive biomass, unlike homogeneous methods. We found that rotational grazing at moderate intensity was the most produc- tive and resilient strategy, while continuous, high-intensity grazing made the ecosystem less resilient. Calibrated for French grasslands, our modular design makes the model a valuable tool for implement- ing spatially explicit grazing strategies in both temperate and arid ecosystems (e.g. African rangelands that deal with unpredictable rainfall and unique pastoral systems).