 This paper compares the suitability of Sentinel-2 and Landsat-8-OLI images for detecting and mapping soil salinity distribution using a deep learning convolutional neural network approach. The study identifies six predisposing variables to train the models, collects ground control points, evaluates different activation, loss of cost, and optimization functions, and finds that Sentinel-2 is more suitable for detecting and mapping SSD than Landsat-8-OLI. The findings demonstrate the effectiveness of the DLCNN approach in supporting fast and reliable image analysis and classification, contributing to understanding, controlling, and managing soil salinization. This article was authored by Mohamed Kazemi-Garajah, Thomas Blashk, Bahid Hosein Hagi, and others.