A comprehensive review of 32 studies (20 journals, 11 proceedings, and one book chapter) published from 2016 to 2023 in the fields of deep learning (DL), image enhancement, super-resolution image, and Geographic Information System (GIS) is presented, focusing on the integration of DL methodologies with GIS to improve the quality of satellite images. The review summarizes the background, principles, enhancement quality, speed, and advantages of these technologies, comparing their performance based on metrics such as Peak Signal-to-Noise Ratio (PSNR), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Structural Similarity Index Measure (SSIM), and computation time. Satellite remote sensing technologies, which have provided an efficient means of gathering spatial information since the launch of Landsat 1 by NASA in 1972, have recently advanced to enable the collection of high-resolution satellite (HRS) images (≤30 cm). However, factors such as atmospheric interference, shadowing, and underutilization of sensor capacity often degrade image quality. To address this, satellite images require enhancement, and DL has emerged as a powerful tool due to its ability to model complex relationships and accurately recover super-resolution images. While DL and neural networks have demonstrated significant success in natural image enhancement, their application to satellite images presents unique challenges. These challenges include insufficient consideration of the distinct characteristics of satellite imagery, such as varying spatial resolutions, sensor noise, and spectral diversity, as well as the reliance on modelling assumptions that may not align with the complexities of satellite data. This highlights the need for further investigation into advanced DL approaches tailored specifically for this domain.