Grape leaf diseases can result in significant yield loss for grape farmers, making accurate and early detection important. This work proposes a deep learning approach using the VGG16 model for grape leaf disease identification. A dataset with 10 different classes served as the basis for the model's training., including 9 disease classes, one healthy class. In this paper, in this study, we evaluate VGG16 and a custom-built CNN for their ability to distinguish between 10 distinct diseases affecting grape leaves. The results showed that both the VGG16 and custom CNN models achieved high accuracy, with a mean accuracy of 77.57% and 78.28%. Respectively. This study highlights the potential of using pre-trained models for grape leaf disease classification and the importance of model selection and fine-tuning for improved performance. This study provides an efficient, reliable, and cost-effective solution for grape farmers to monitor their crops and prevent the spread of diseases, ultimately improving crop yields.