This research presented in the paper revolves around the classification of kyphosis disease, a spinal condition characterized by an abnormal curvature of the upper spine, leading to a rounded back. The primary objective is to develop a predictive model that can accurately determine whether a patient has kyphosis based on specific diagnostic measurements. Exploratory Data Analysis was employed to conduct preprocessing tasks with one hot encoding. By investigating machine learning and deep learning algorithms, particularly decision trees and random forest and CNN, this research aims to achieve a high level of accuracy in predicting the presence or absence of kyphosis. Overall, the research contributes to the broader understanding of kyphosis disease classification and highlights the potential of AI-driven techniques in medical diagnostics. Model performance was evaluated by Confusion matrix, precision, Recall, F1 score, and Accuracy. The achieved accuracy level of 93.33% for CNN demonstrates the efficacy of the proposed approach and its relevance in clinical practice for diagnosing and managing kyphosis effectively.