A key component of international agriculture, potato farming, provides the majority of the food for many communities around the globe. However, a number of illnesses often endanger the well-being and yield of potato crops, which can cause significant financial losses. Sustainability in agriculture aims to meet current food needs without harming future generations’ ability to meet theirs. It involves practices that conserve resources, protect the environment, and maintain soil health. The goal of this research is to investigate YOLOv5 architecture to identify illnesses in potato plants. A dataset of 1560 real-time images of potato plants suffering from various infections were collected. Investigation has been performed on three distinct deep learning architectures YOLOv3, YOLOv4, and YOLOv5 for detecting diseases in potato crops. Among these, YOLOv5 produced superior outcomes than its previous variants. YOLOv5 achieved precision of 0.97 for the healthy plants and 0.96 for the diseased plants. By utilizing deep learning models YOLOv5, it is possible to improve food security and agricultural sustainability while also promoting sustainable and ecologically friendly farming practices and early and precise detection of potato diseases.