Agricultural production in "open-sky systems" such as extensive cattle ranching on natural grasslands is subject to inter-annual climatic variability and other market fluctuations. To tackle the dependency of livestock breeding on these factors, we conducted participatory modeling with cattle producers in Uruguay. The methodology consists of simulating possible scenarios to collectively evaluate the different herd management practice alternatives. In this paper, we present an Agent-Based Model built with stakeholders and designed to represent a breeding system on a typical extensive grazing area in the basaltic soils region (BR) of Uruguay. This model has three main modules: environment, biophysical and decisional sub-models. This modularity allows the conducting of virtual experiments to reveal how some herd management decisions (such as seasonal stocking rate adjustments) combined with a climatic series can result in resilience against drought periods and market movements. Long-term simulations were implemented to analyze the sensitivity of the model to key management parameters with varying climate conditions. The inter-annual climatic variability can seriously affect cattle production, even with conservative stocking rates. Rigid strategies are bound to fail and cause systems to break. Adaptive management emerged as a critical option for the sustainability of livestock breeding. The inter-annual climatic variability can seriously affect cattle production, even with conservative stocking rates. This result highlights the importance of adaptive management, one that can react to a changing environment, for the sustainability of livestock breeding.