<p>The rising demand for sustainable farming underscores the need for early plant disease detection, especially in vegetables. Early action prevents crop losses, boosts food security, and supports eco-friendly practices. This study such as vegetable leaf disease detection and recognition (VLD2R) presents a simple, scalable framework using a lightweight, pretrained deep learning model to spot vegetable leaf diseases early. This approach combines advanced feature extraction and optimization for reliable disease detection using a pretrained model. A fine-tuned, compact convolutional neural network analyzes diverse leaf images to detect common issues accurately. Leaf classification, vital for agriculture, is tough due to shape variety and environmental changes. The system VLD2R enhances image quality through pre-processing, then uses the pretrained model to identify fungal, bacterial, and viral diseases. The overall accuracy across four data sets: pepper bell (dataset I) at 99%, potato (dataset II) at 99. 2%, pumpkin (dataset III) at 88%, and tomato (dataset IV) at 98. 48%. The model excels in pepper bell, potato, and tomato datasets, with precisions greater than 98%, but its performance slightly differs for pumpkin at 88%, likely due to the variability in class scores. Tested on a large vegetable image dataset, it excels at early detection, even with limited data. This practical tool helps farmers monitor crops, apply targeted treatments, reduce pesticide use, and lessen environmental impact. By improving yields and adapting to various crops and regions, this affordable, green solution advances global precision agriculture.</p>