The main traditional livelihood in the high Central Andes Plateau, locally known as the Puna, is pastoralism. This system is based on llamas (Lama glama) and sheep (Ovis aries), which graze on natural vegetation (shrub steppe, grasslands, and peatlands). So, the livelihood of Andean communities is strongly influenced by the spatial and temporal dynamics of vegetation. Most of the existing models on herbivory used to estimate sustainable livestock loads require information on the nutritional values and primary production of vegetation, which is lacking for the Argentine Puna. This work aimed to fill the gap by implementing a carrying capacity model using data on annual vegetation growth and herbivore feeding strategies collected in the field in different plant communities. The hypothesis that guided this work was that environmental heterogeneity affects management decisions and should be taken into account to model sustainable animal stocks in Andean pastoralist systems. We estimated carrying capacity based on the MSY logistic model, considering biomass and aboveground primary production of vegetation communities, climatic features, and herbivore selective responses. We compared the estimated carrying capacity with real animal stocks in four production units within the study site in 2013, 2014, and 2017. The results indicate that the carrying capacity presents high spatial and temporal variability. In three of the production units assessed, the livestock—mixed herds of sheep and llamas—were above or near the estimated carrying capacity, even when the stocking rates were reduced during dry periods. The high stocking rates, along with evidence of land degradation, may indicate moderate to severe overgrazing. This work demonstrates the usefulness of estimating the carrying capacity to monitor the pastoral systems at fine space and time scales. These monitoring techniques can facilitate decision making among Puna herders in a biocultural context of knowledge dialog that may foster sustainability in pastoralist livelihoods.