The improper disposal of cow dung poses severe environmental and health challenges, including the spread of vector-borne diseases and respiratory issues. Conventional methods for dung collection, reliant on manual labour and traditional tools, are inefficient, labour-intensive, and expose individuals to significant health risks. This study addresses these concerns by developing an automated cow dung detection and collection robot that combines advanced AI, machine learning, and robotics to provide an efficient, safe, and cost-effective solution. The proposed system utilizes the YOLOv8 deep learning model for real-time detection of cow dung with an accuracy of 85%. A dataset of 200 images, covering various natural environments, was meticulously pre-processed, annotated, and used for training the model. The robot’s mechanical framework employs a forklift mechanism and a lead screw for efficient dung collection, guided by a Raspberry Pi and Arduino Nano, which coordinate the detection and collection processes. The system is capable of collecting up to 150 kg of dung per day at an operational cost of 1 rupee per kilogram, demonstrating its practicality and scalability for real-world applications. This research highlights the robot’s superiority over conventional systems in terms of health safety, cost-efficiency, and environmental impact. The design and performance validation indicates that the robot is a sustainable solution for rural and urban waste management.