Potato farming is crucial for worldwide agriculture, but illnesses provide substantial risks to both crop productivity and quality. This proposed work presents a novel approach that combines Convolutional Neural Networks (CNNs) with sophisticated algorithms such as Support Vector Machines (SVM) to detect diseases in potato leaves. The dataset is divided into three parts: training, validation, and testing. The training set consists of 80% of the data, while the validation and testing sets each contain 10%. The dataset contains a total of 13,222 files, which include samples from the healthy, late blight, and early blight classes. These files were sourced from published work and a few from the SRM Institute of Science and Technology Agriculture field. After running 25 epochs, the system achieves an accuracy of 97.01%. Through the analysis of leaf images, the presence of diseases and evaluate the level of confidence in the classifications. The integration of Support Vector Machines (SVM) improves accuracy. SVM is particularly effective at identifying the optimal separation between classes. Accuracy, recall, and F1-score are computed for proposed model respectively. This integrated methodology merges deep learning with sophisticated algorithms, enhancing the precision of disease diagnosis and enabling efficient agricultural practices for the sustainable development of potatoes on a global scale.