We propose a combination of methods for processing multitemporal multispectral satellite images to identify variations in vegetation indices for fixed regions across time to analyze deforestation in Peru. The images, taken from a location in Ucayali, Peru, are obtained from public repositories for Sentinel-2 satellite, which covers up to 13 spectral band layers with spatial resolutions of 10, 20, and 60 m and have a temporal resolution of 5 days with the same angle vision. We preprocessed the images for cloud and river removal and normalization to apply the normalized difference vegetation index (NDVI). All images were segmented with their corresponding model, leading to the analysis of the variance of this segmentation information across different time frames for a single location in which the normalized difference of time series (NDTS) was obtained for each year. Finally, a first approach was built for a linear regression model that could segment each frame into sectors, according to vegetation density, using the multispectral features as parameters. For the selected location, the results show an increment from about 4200 ha in 2018 to about 8200 in 2020.