Drought detection is essential for effective environmental and agricultural planning. In this study, we leveraged satellite imagery from the USGS Landsat8 dataset, focusing on RGB bands, to develop a deep learning-based model for identifying drought conditions. To classify the data, we used the number of cows present in the images—images with 0 or 1 cow were labeled as Class 0, while those with 2 or more cows were categorized as Class 1. Drought conditions are detected if an image falls under Class 0, whereas Class 1 indicates no drought. We trained and evaluated four deep learning models: EfficientNetB0, MobileNetV2, VGG16, and a custom CNN. By analyzing performance metrics such as accuracy, precision, recall, and F1-score, we found that EfficientNetB0 outperformed the other models, with an accuracy of 90.31% and a precision of 90.58%. These results demonstrate its reliability in detecting drought-prone areas. Our findings emphasize the potential of deep learning for satellite-based drought detection, offering valuable insights for environmental monitoring and resource management.