Agriculture is India’s largest livelihood provider and contributes to the GDP. India has the world’s second-largest agricultural land, generating employment for a large population. Using a leaf image dataset, this paper uses machine learning and optimization techniques to detect maize plant disease. This paper introduces the new path for early stage disease detection using a maize leaf image dataset. This work proposes a modified Grey Wolf Optimizer (MGWO) for disease identification and classification based on the maize leaf image dataset. This method identifies the particular disease for early stage treatment. The modified GWO is trained and tested with the help of four different categories of maize leaves, including blight leaf, common rust leaf, gray leaf spot, and healthy leaf, and distinguishes diseased and healthy leaves. This method used the random forest classifier to classify the images into different categories, find the accuracy, and compare it with the accuracy of the existing algorithms. The comparison shows a better result for the proposed algorithm than the current algorithms.