Accidental oil spills are known to reflect negative outcomes on the environment and human health as well as marine life and coastal regions’ economy. In aim to solve this issue, we suggest a system that is designed to detect oil spills on the ocean surface and provide information for taking appropriate measures to contain the spill. Our research focuses on two key maritime regions near the United Arab Emirates: The Arabian Gulf and the Gulf of Oman. To create the dataset, we utilized Sentinel-1 Synthetic Aperture Radar (SAR) images that were pre-processed using SNAP for training and SNAPPY in Python for testing. The system uses an automated Vision Transformer (ViT) as its base for the classification and segmentation of oil spills, which was trained on SAR patches falling under two classes. The step of automating the system involves receiving new data inputs and outputting image segments containing the oil slick without delay. Our proposed approach illustrated a high level of performance compared to other Convolutional Neural Network architectures used in similar cases. The ViT accomplished 0.91 accuracy on unseen data with error of 0.3. We put the model into test on new SAR images. The suggested system will help minimizing the effects of oil spills on the ecosystem, human health, and economic losses in the UAE. We believe this study will mark a breakthrough in the management of oil spills as it seeks to safeguard crucial marine and coastal resources through engaging Artificial Intelligence (AI) with cutting-edge algorithms.