Chronic Kidney Disease (CKD) is major concern of death in recent years that can be cured by early treatment and proper supervision. But early detection of CKD and exact risk factors should be known to ensure proper treatment. The study mainly aims to address the issue by building a predictive model and discovers the most significant risk factors employing machine learning (ML) approach for CKD patients. Four individual machine learning classifiers were applied to conduct this study. It is found that GB performed very poor compare to other applied classifiers where RF and LightGBM outperformed with 99.167% accuracy. In terms of risk factors, it is found that sg, hemo, sc, pcv, al, rbcc, htn, dm, bgr, and sod are the most significant factors, which are mainly correlated with CKD. The study and its findings indicate that it will enable patients, doctors and clinicians to identify CKD patients early and ensure proper treatment for them.