Oil spills in the ocean are causing serious damage; therefore, technologies and tools for oil pollution detection should be improved quickly. This research analyzes current progress that has been made in machine learning (ML) and deep learning (DL) technologies by focusing on their ability to detect oil leaks. After laboriously going over numerous studies, we’ve identified trending designs of neural networks that can be used to precisely detect and classify oil spills. This isn’t just about what works though. We also point out obstacles such as telling an oil spill apart from its almost indistinguishable natural twins or having few datasets, leave alone unavailability of enough datasets while having uncertain patterns. Wrapping up, we chart out the path forward, focusing on how to translate these theoretical marvels into real-world saviors for immediate oil spill surveillance and action. Through this exhaustive review, we’re keen on enriching the dialogue and setting the stage for future breakthroughs in ML and DL that could be game-changers in how we monitor and safeguard our marine environments from the scourge of oil spills.