Abstract: In the statistical analysis of clinical data, we often encounter the situation of missing data. According to the condition of missing data, proper methods to deal with missing values are related to the reliability and accuracy of statistical inference. In this paper, the model of missing data, proportion of missing values, and the common processing method were briefly summarized; the process and content of multiple imputation based on R package mice were shown in detail. Finally, an analysis of the influencing factors of acute kidney injury was conducted to illustrate multiple imputation, and the results of generalized linear models based on multiple imputation data sets, missing data sets, and original data sets were presented and compared. The results showed that multiple imputation did well in the imputation of quantitative variables, dichotomous variables and ordered multiclassification variables, and the result of the model based on multiple imputation data sets was robust and reliable.