Chronic kidney disease (CKD) is a progressive disorder that worsens over time, leading to a variety of major health problems including hypertension, anaemia, nerve damage, fractured bones, and cardiovascular diseases. This rate is growing globally by the development of ageing and diabetes and hypertension. Precise prediction of CKD may aid healthcare practitioners in developing tailored treatment plans that address the distinct underlying causes of the illness, reduce the likelihood of complications, and improve patient outcomes. Predicting the course of CKD is crucial because early detection and accurate diagnosis of CKD may improve patient outcomes, prevent complications, and save medical costs. Machine learning (ML) based techniques are taken into consideration for CKD prediction. Several ML algorithms are used, including decision tree (DT), neural networks (NN), random forest (RF), XGBoost (XGB), Gaussian Naive Bayes (GNB), and CatBoost (CB) Classifier. Using six distinct ML algorithms, the kidney disease dataset is used to carry out the empirical method. Different validity ratings are used to evaluate and contrast the generated findings. The results show RF, XGB and CB get higher accuracy of 95 % which is better suited in predicting CKD than DT, NN, and GNB. These algorithms illustrate the highest accuracy and efficiency in the diagnosis of the patients at the beginning CKD stage. An accurate prediction of CKD enables practitioners in the healthcare sector to develop individualized treatment plan, avert the occurrence of complications and achieve a better patient outcome. Early recognition and intervention based on predictive models can prevent unproductive testing, treatments, and hospital beds while also responding quickly to disease management. The predictive models of CKD can identify high-risk patients and facilitate early interventions, ultimately enhancing public health outcomes.