Forests are essential providers of oxygen, playing a key part in maintaining life. They serve as crucial habitats for a wide range of species, promoting biodiversity. The change in the forest region will cause a severe impact in the ecological equilibrium. Satellite image (SI) supported forest area monitoring is one of the prime practices to analyze the increase/decrease in the forest area. This research aims to develop an automatic segmentation procedure to extract and evaluate the forest region using a deep learning-based scheme (DL-scheme). The stages in the proposed scheme involves; (i) image and mask collection and resizing; (ii) enhancing the image using Kapur’s entropy and Firefly Algorithm (KE + FA), (iii) extracting the forest section with ResUNet model, and (iv) comparing it with the mask and computing the necessary image metrics. The experimental investigation is executed using the Matlab- and Python-software and the obtained outcome with the ResUNet is confirmed against the UNet, UNet + , and UNet +  + . This study provides a segmentation accuracy of > 98% on the chosen database when KE + FA enhanced images are considered.