The forest is called the green lung of the planet. Unfortunately, this sentence is forgotten day by day and various types of pollution do not help on this matter, especially the one that affects it directly - deforestation. To slow down this phenomenon, it can be used the technological advancements in the recent years in conjunction with machine learning techniques. This paper proposes a method for this matter, by applying an agglomerative clustering algorithm to a set of aerial images and analyzing different methods of metrics and linkage in the process, to monitor deforested areas. The study showed that Ward's method of the AGNES algorithm had the best results, even better than a previous k-means approach implemented on the same dataset.