This study investigates the use of the YOLOv8 deep learning model for automating the detection of farm ponds in the rural environment of Maharashtra, India. Agriculture is a key sector in Maharashtra, facing challenges like irregular rainfall and soil degradation. Farm ponds, artificial structures for water storage, are promoted as a solution to these issues. To detect these ponds, satellite imagery from Jalna district is processed by dividing the images into smaller tiles, adding bounding boxes, and preparing data for model training. The YOLOv8 model is trained using these images, and the performance is evaluated based on key metrics like mean average precision (mAP50) and detection confidence. The model achieves an overall accuracy of 85.85%, with 86.30% accuracy for detecting wet ponds and 84.85% for dry ponds. The study demonstrates YOLOv8's effectiveness in identifying farm ponds in satellite imagery, with strong results in both wet and dry conditions. This automated approach offers a scalable solution for monitoring farm ponds, essential for water management in agriculture. By using deep learning to detect farm ponds, the method provides a tool for proactive measures in drought mitigation and sustainable farming practices, enhancing water conservation and crop planning. The success of YOLOv8 in this context suggests its potential for broader applications in agricultural monitoring and resource years.