Sound management of native forests used for cattle grazing requires understanding the dynamics of forage productivity in the openings. Despite their importance, forage productivity drivers in highly heterogeneous forested landscapes, or their variability over the year, are still unclear. The aim of this work is to find predictors of Normalized Difference Vegetation Index (NDVI) variation in the openings of native temperate forests and to evaluate how these predictors change within the growing season. We used high spatial resolution remote sensing imagery from NW Patagonia to separate forest openings from tree dense canopy. We obtained data of each opening related with herbaceous and shrub forage productivity and calculated landscape metrics. We estimated a multiple linear regression model for predicting NDVI in each season. Beyond known variables related with forage productivity (altitude, precipitation, etc.), the shape of forest' openings appeared as relevant in predicting NDVI. Higher values of forest opening perimeters were related with a decrease in NDVI in spring when soil water content is not limiting and conversely with an increase in NDVI in summer when water is limiting growth. These results suggest that environmental drivers such as temperature and soil moisture inside the opening, and competition or facilitation process between trees and grasses are mediated by the shape of the opening. Management of heterogeneous native forests for cattle raising requires considering the shape of the openings to maximize forage productivity.