Chronic kidney disease (CKD) is a global health problem with high mortality and morbidity and mortality. Real-time performance using machine learning. In this study, we introduce machine learning for CKD diagnosis. CKD data is from the University of California, Irvine (UCI) Machine Learning Repository, which contains many missing values. KNN assignment selects multiple completed models with the best values ​​to predict missing data for each incomplete model and is used to load missing values.Although patients may ignore certain measures for a variety of reasons, missing data is often found in real clinical settings. After solving the missing data, models are constructed using machine learning algorithms (logistic regression, random forest, support vector machine, k-nearest neighbor, Naive Bayesian classifier, and feedforward neural network). Random forest machine learning models are the most accurate in this task.