Clouds play a pivotal role in the Earth’s climate system, significantly influencing climate dynamics and climate change. Accurate identification and analysis of cloud cover are essential for meteorological research and satellite remote sensing, providing critical insights into climate variability. Despite its importance, cloud detection is challenged by data noise and external factors like ice and snow. Recent advances in artificial intelligence, particularly deep learning techniques, offer promising solutions to these challenges. This research leverages state-of-the-art deep learning models, including UNet, UNet++, ResUNet, and Attention UNet, for feature extraction and semantic segmentation in cloud detection. By generating precise cloud masks that separate clouds from their background, these models outperform traditional methods in accuracy and robustness. The study also underscores the versatility of deep learning in processing optical images beyond cloud detection. In summary, this work integrates climate science with advanced deep learning models, creating a powerful tool for meteorologists, climatologists, and remote sensing experts. The use of metrics like the Jaccard Index and Dice coefficient ensures a comprehensive evaluation of model performance. This research not only enhances our understanding of climate dynamics but also sets a precedent for innovative applications in image processing.