The issue of missing values in the scholarship selection process poses a challenge that can impact decision-making. This study aims to perform data imputation for scholarship candidate datasets using the K-Means method and evaluate its performance using the Mean Absolute Percentage Error (MAPE). K-Means was selected for its ability to group data based on pattern similarities, enabling it to estimate missing values in the scholarship candidate dataset. Two datasets were utilized in this study: one with 10% missing data and another with 20%. The results indicate that K-Means imputation can effectively apply to scholarship candidate data. Additionally, the findings reveal that the proportion of missing data influences the optimal number of clusters required. For the dataset with 10% missing data, the best configuration was achieved with 5 clusters, resulting in a MAPE of 13%. Conversely, for the dataset with 20% missing data, the optimal configuration required 2 clusters, yielding a MAPE of 14%.