Chronic kidney disease (CKD), due to rising patients, the high probability of deterioration towards end-stage renal disease and inaccurate estimates of morbidity and mortality, constitutes a heavy burden on the sanitary infrastructure. The aim of this research is to build a model for machine-learning, which uses comorbidity and data on drugs and predicts population prevalence. Predictive health care prediction using machine learning is a daunting activity to help clinicians assess the precise therapies for life-saving. In this paper, the study applies machine learning method in combination with ensemble learning for estimation of chronic kidney disease with clinical evidence. They are based on chronic kidney disease datasets and the efficiency of these models is compared to choose the best classifier for chronic kidney disease prediction. The comparative analysis is estimated in terms of various metrics like classification accuracy, f-measure, percentage error, etc. The results of simulation shows that the proposed ensemble machine learning classifier namely Ensemble Support Vector Machine predicts well the chronic kidney disease from the datasets than other existing ensemble methods.