By 2020, almost all countries in the world will face the COVID-19 outbreak, including Indonesia. One of the impacts that occurred due to the COVID-19 pandemic was the obstruction of statistical activities, such as delayed or stopped carrying out survey and census data collection and other data collection. Meanwhile, to meet data demands and needs during the COVID-19 pandemic, the national statistical agencies must continue to collect data and provide statistical data. Therefore, the national statistical agency must adapt to the census and survey process activities carried out, such as finding alternative data collection modes, reducing sample sizes, modifying sample designs, reducing question items in questionnaires, or others. Based on this description, the adaptation of census / survey data collection activities carried out during the COVID-19 pandemic will affect the quality of the data produced. One of them is missing data. To solve the problem of missing data, one method that can be used is data imputation. One type of machine learning-based imputation method that is often used is Weighted K-Nearest Neighbor Imputation (Weighted KNNI). The Weighted KNNI method has better accuracy than the other two imputation methods (Unweighted KNNI and Mean Imputation) for each percentage of missing data, both the accuracy from the RMSE side and the accuracy from the MAPE side. Based on these results, seen from its accuracy, the KNNI Weighted method can be used as a solution to dealing with incomplete data during the current COVID19 pandemic