In Asia, rice is the most popular staple food for almost half of the world’s population. Increasing rice production wherever possible is the least expensive strategy to keep up with population expansion. Paddy leaf infections are one of the main issues in rice production. Therefore, an automated approach for identifying and categorizing paddy leaf diseases is suggested by considering current trends and difficulties. This system uses a dataset from Mendeley Data that contains four different types of rice leaf diseases: Bacterial Blight, Blast, Brown Spot, and Tungro.This system consists of three sub- processes that are pre-processing, feature extraction and classification. Grayscale images are created from RGB photos during the pre-processing stage. The grayscale image is then cleaned of noise using the median filtering technique. The OTSU thresholding approach is used to segment the diseased and non-diseased portions. The segmented image's features are extracted using the gray-level co-occurrence matrix technique. This method uses the AlexNet deep learning model to classify the paddy leaf disease using the retrieved features.