Native forests around the world are widely used for livestock grazing as they offer different sources of forage. Nevertheless, in heterogeneous forested landscapes, forage productivity drivers are still unclear, hindering precise predictions of field receptivity. Our aim is to relate landscape variables with forage productivity in forested landscapes using satellite and ground-based data. To accomplish this, we harvested 36 enclosures in two Patagonian valleys during a three-year period. The location of the enclosures encompassed a gradient of altitude and mean annual rainfall, across three vegetation types commonly used for cattle raising. We estimated five generalized linear models to predict forage productivity using different remote sensing and ground (field) data as predictors. The most important variables for predicting forage productivity were five of remote sensing type and two of field type. The highest goodness of fit was obtained when all variables were included (D2¿=¿0.71). When ground-based information was combined with remote sensing data, the goodness of fit was higher (D2¿=¿0.65) compared with models that only used remote data as predictors (D2¿=¿0.49). Models obtained based on remote data serve to support situations where field information is not available. High forage productivity levels can be obtained in high forests or scrubs with varying values of canopy openness, without removing the forest. The models generated in this work improve livestock stocking rates adjustment in NW Patagonia forests, and may be also re-estimated with new data in other regions used for cattle raising worldwide, contributing to the sustainable use of native forests.