Livelihoods are being impacted by the danger posed to water supplies by changing climates, growth in the population, and degradation of the land. Observing the surface water bodies is essential for evaluating historical, current, and anticipated changes. Satellite imaging and altimetry are used to assess fluctuations in water levels and accumulation in tiny water bodies. This study introduces a comprehensive methodology for the detection of water bodies in satellite imagery utilizing the U-Net and VGG-16 architecture. Through rigorous preprocessing of datasets, the U-Net model is trained to accurately delineate water bodies across various environmental contexts. Data preprocessing encompasses tasks such as image resizing, normalization of pixel values, and annotation of water bodies, ensuring the precision of model training. Employing carefully calibrated hyperparameters, the model undergoes training with meticulous monitoring of its performance. The findings demonstrate the U-Net and VGG-16 model’s effectiveness in accurately detecting water bodies, indicating its potential for applications in environmental surveillance and management.