India is the world's second-largest consumer of cashew nuts. Cashew business is also responsible for the livelihood of many farmers. The potential for cashew nut production has not been realized due to some obstacles such as traditional cultivation practices, extreme weather conditions, diseases and pest prevalence. As these nuts are very limited in production, so they must be identified early for effective management. This research focuses on the problem of the leaf blight of cashew plant using advanced machine learning and deep learning algorithms. Five algorithms are evaluated to determine the best model, those are Support Vector Machine (SVM), Random Forest, MobileNet, InceptionV3, ResNet152V2. The research methodology involves data preparation, model training and results evaluation, demonstrating the effectiveness of Convolutional Neural Network (CNN) models for early disease diagnosis. The proposed approach offers reliable and effective method for early disease identification of cashew plants, which can lead to increased yields and better agricultural practices.