Detection of paddy leaf disease is crucial for the agriculture industry since rice provides sustenance for over 50% of the global populace. In this paper, a novel YOLO-DBN framework has been proposed to identify the leaf diseases like blight, smut and spot in paddy crops. The paddy leaf images are pre-processed using CLAHE (Contrast Limited Adaptive Histogram Equalization) to increase the quality of the images. The pre-processed images are fed as input to YOLO Network to conduct instant segmentation of paddy leaves. The segmented images are fed as input to Deep Belief Network to classify the paddy leaves into blight, smut and spot diseases. The proposed YOLO-DBN achieves a high accuracy range of 97.68%, 96.71% and 98.76% for detecting Blight, smut and spot respectively. The proposed approach and the conventional deep learning techniques like DNN and Alex net. A clustering algorithm is utilized to segment the backdrop, normal section, and sick region. The proposed YOLO-DBN model improves the overall accuracy of 7.51%,1.18%, and 0.38 % better than CNN, DNN, Alex net respectively.