The presented work evaluates the potential of a Sentinel-2 time-series to detect Pinus radiata (Monterey Pine) invasions in endemic Nothofagus (Southern Beeches) forests in the Maule region, central Chile. Suitable cloud free images of the phenological cycle were selected from six Sentinel-2 scenes available for the years 2016. The scenes were unmixed using a non-negative least square (nnls) algorithm for different landcover components per pixel. The results were validated with a SVM classification of UAV-based hyperspectral mosaics acquired in March 2016. The results show that it is possible to map the coverage of the Pinus radiata class up to an accuracy of R2 0.6 and RMSE of 10 %. Generally, a higher number of Sentinel-2 images increased the performance of the model, while there was no significant dependency on a specific acquisition date. However, the variability of the results is high, which indicates that a careful selection of multi-temporal endmembers is crucial to a successful unmixing of Pinus radiata.