Pea leaf diseases constitute a suggestive prospect to crop yield and quality, making early and precise detection essential for effective management. This study investigates advanced machine learning (ML) and deep learning (DL) techniques to point out and grade diseases in pea leaves. To outstrip classification exactness, we propose a hybrid model that integrates EfficientNetB7, DenseNet121, and MobileNet architectures, leveraging the peculiar advantages of each to capture diverse features within leaf images. Our dataset includes an even amalgamate of healthy and diseased pea leaf images, dissect into training, validation, and testing sets for accurate model estimation. The purpose is to create a highly accurate and adaptive solution for detecting pea leaf diseases, promoting sustainable and efficient farming practices. The model is staging exceptionally well in distinguishing healthy leaves from diseased ones. Unlike traditional manual methods, this automated approach provides high accuracy, making it suitable for large-scale agricultural use. By using data amplification and remodel, the replica becomes more malleable to different environments and disease types. Prompt results are propitious, with accuracy approaching 100%. This inspection supports smart farming by handing out a reliable tool that helps farmers act quickly to turn down garnering losses and manage endow better.