The aim of this study was to evaluate the potential of MODIS normalized difference vegetation index hypertemporal data analysis for assessing Mediterranean pasture conditions in North Western Sardinia (Italy). During the seasons 2006 to 2007 and 2007 to 2008, field observations were carried out to classify 67 pasture sites in three condition classes based on expert knowledge. The local net primary productivity scaling (LNS) method was applied, and its potential for discriminating the pasture condition classes was evaluated by logistic regression models (LRM). Yearly and average LNS maps were generated for the period 2000 to 2008, and analyzed to identify areas that exhibited persistently low LNS values (hotspots). The LNS method proved useful to discriminate pastures in different conditions (LRM bootstrapped Nagelkerke pseudo ). The analysis of persistence of low LNS values allow identifying regional hotspots of degradation. A qualitative evaluation of the main hotspots on aerial photographs revealed that approximately 62% of the hotspots were clearly characterized by pasture degradation patterns, whereas the remaining were associated to highly fragmented landscapes or to errors in the land cover map. This result emphasizes the importance of using multiscale approaches by integrating the LNS regional assessment with high spatial resolution remote sensing data analysis.