Agriculture plays an important role in everybody’s daily life, from farmers to consumers. Humans mainly depend on plants for their diet. Early plant leaf disease prediction is crucial to ensure sustainable crop production and less wastage of diseased crops. It helps save time, money, and resources in yielding healthy crops. Deep learning can play an important role in helping agriculture detect early diseases in leaves to prevent the entire plants from dying. This study delves deeper into LeafGuard, which is a custom deep learning model built on the principles of ResNet-18 and Convolutional Neural Network (CNN). LeafGuard helps identify diseased leaves from healthy ones. This model classifies a leaf image based on thirty-eight classes ranging from different species and types of diseases. The proposed model has reached an accuracy of 97.55%. LeafGuard can significantly help reduce the time and resources it takes farmers to manually look for diseased leaves. Diseased crops can damage soil, water, and the environment and reduce the health of crop fields for future use. With the help of LeafGuard, sustainable development can be achieved. This is because early diagnosis of diseases leads to reduced crop losses, optimized resource use, and minimized environmental impact. In this way, we can achieve sustainable development, so that our future generations will have more crops reserved.