This work presents productive and reproductive results of the simulations made with an extensive livestock farm model on basaltic soils. The scenari were constructed varying the initial grass height (GH) and the system stocking rate (SR). The model is chaotic, dynamic, empirical and deterministic. It was built in a participatory way with a group of local stakeholders of the basaltic region of the country (Uruguay). A simplified version of it was utilized, where only one parcel and only one animal category were simulated each time. No adaptive management was implemented into the simulation process. These models represent a Breeding (Br) and a Rearing (Re) extensive systems, where the exclusive feed source is the natural pasture. All scenari were simulated with average pasture growth, estimated by satellite remote sensing for the series 2000-2010, for a representative farm basaltic region. Annual results of pregnancy rate and live weight gain per animal were the two main outputs analyzed for the Br and Re simulated systems, respectively. In the scenari building the two input variables changed stepwise were the initial GH (going from 2 to 12 cm) as indicator of the forage availability, and the initial number of animals (going from 0.1 to 1.9 animal/ha) as indicator of the SR. The initial situation for the Br simulation was cows of 360 kg of average live weight. The start of the simulation was at the beginning of autumn, and the calving and mating seasons were spring and summer, respectively. For the Re simulation, the initial situation was defined by heifers with an average live weight of 150 kg. Simulation started also at the beginning of autumn. For the Br simulations results of pregnancy rates higher than 80% are only achieved in the scenari of low and mean SR (<1.1 cows/ha) and high GH (> 10 cm). If the initial GH on autumn is 5 cm, to achieve that pregnancy rate level, the SR must be less than 0.9 cows/ha. For the Re simulation, results of gain per animal bi