Resumo. In general, the Mediterranean forests have suffered high levels of fragmentation, mainly by anthropogenic action, as result of the productive activities of agriculture and livestock, human settlements and forest plantations of fast growing exotic species, leading to loss of continuity of natural forests, and thus a decrease of biodiversity. In this context, the objective of this study was to evaluate the potential use of satellite images and Geographic Information Systems (GIS) to quantify and characterize the spatial evolution of native forests in a semi-arid landscape area at the Region of Maule, Chile. The study was conducted with Landsat-5 TM images of 2004 and 2008, which were classified by the supervised method where it was previously necessary to apply the atmospheric and the geometric corrections to reduce errors of interpretation and measurement. Within the analysis, it was incorporated digital thematic information of land use of the National Forestry Corporation (CONAF) and also it was joined one of the latest techniques derived from remote sensing for studying changes in land cover, that is called Radiometric Rotation Controlled by No-Change Axis (RCNA). Finally, the fragmentation indexes were estimated for characterizing the spatial configuration of landscape. From the results of the digital classification it was possible to estimate landscape metrics that show significant differences in the structure of the study area between the two periods. In addition, it was generated an image of change 2004/2008 that allowed to evaluate the consistency of the RCEN technique, demonstrating the feasibility of using this alternative methodology for the detection of changes in terms of degradation and recovery of woodlands.