Intelligent Agriculture: Plant Disease Detection, Crop Prediction, and Fertilizer Recommendation Using Machine Learning

Agriculture’s fundamental mission is to sustain human life, yet modern farming continues to face pressing challenges such as declining productivity, unsustainable resource use, and environmental degradation. As the demand for sustainable practices grows, there is an urgent need for innovative, data-driven solutions that can empower farmers and optimize agricultural outputs. This study presents an integrated machine learning-based framework that supports three key agricultural functions: crop selection, fertilizer recommendation, and early plant disease detection. The system provides personalized crop and fertilizer advice by analyzing local soil characteristics—namely nitrogen, phosphorus, potassium levels, pH, and rainfall—to deliver regionally optimized recommendations. Additionally, the framework incorporates a Convolutional Neural Network (CNN) to detect and classify plant diseases from leaf images, offering actionable treatment suggestions. Publicly available datasets from Kaggle were employed for training and evaluation, including 87,867 images for disease detection, 4,540 for crop classification, and a dataset with 4,513 soil records for fertilizer advice. To ensure accessibility, the system is delivered through a user-friendly web interface, enabling even non-technical users to leverage advanced agricultural analytics. By enhancing yield, minimizing crop loss, and promoting efficient use of resources, this framework contributes to sustainable farming practices. Experimental results demonstrate high accuracy, precision, and recall, highlighting the system’s potential to support resilient and eco-friendly agriculture