Marine heatwaves are periods of unusually warm ocean temperatures that can harm fish, ecosystems, and people who depend on the ocean. In recent years, these events have become more frequent and intense, especially along the coast of Peru, one of the richest fishing areas in the world. However, most global climate models cannot represent important local features, like coastal winds and sea temperaturas, because they work at very large scales. In this study, we combine two techniques, machine learning and ocean modeling, to improve climate projections at regional scales. First, we use artificial intelligence (specifically, convolutional neural networks) to improve the resolution of wind data from global models. Then, we use those winds to run a high-resolution ocean model, which gives more detailed sea surface temperatures. These results are then used to train a second machine learning model to expand the analysis to more climate models. This hybrid approach helps us better understand how marine heatwaves could change in the future. The information can support coastal communities and fisheries by helping them prepare for and adapt to changing ocean conditions.