The use of remote sensing and digital processing of images is increasingly present in the monitoring and analysis of environmental data, mainly due to the ease of obtaining orbital images and the advanced methods of image classification. Temporal evaluation of restoration of native vegetation surrounding springs is important for the protection of water resources. A project named Water Conserver (Conservador das Águas) was developed between 2005 and 2015, involving 380 springs in Extrema, state of Minas Gerais, Brazil, focused on the preservation of water sources. The present study was carried out to quantify and analyze Permanent Preservation Areas (PPA) surrounding these water bodies. Free orbital images with spatial resolution of 30 m were acquired from the satellites Landsat-7 and Landsat-8 in 2003 and 2021, respectively. Object-oriented classification was used to evaluate the total PPA of restored springs over 18 years. The success of the Water Conserver Project resulted in a total restoration of more than 82% of areas surrounding springs up to 2021.