Pregnant women all over the globe face the serious health risk of gestational diabetes mellitus (GDM). A type of glucose intolerance, it is characterized by the development or discovery of high blood sugar levels during pregnancy. Not only does gestational diabetes mellitus (GDM) affect the mother's health, but it also puts the unborn child at risk of complications like macrosomia, birth defects, and the need for a cesarean section. The aforementioned issues can be addressed with the aid of machine learning. Hot Deck Imputation consistently outperforms No Imputation across all models tested (Bayes Net, Decision Table, IBK, Multi-Layer Perceptron, and Random Forest). The accuracy for Hot Deck Imputation ranges from 95.97% to 96.97%, while No Imputation accuracy ranges from 78.33% to 83.27%. This substantial difference is also reflected in other metrics such as Precision, Recall, ROC, and PRC, where Hot Deck Imputation shows higher values. The MLP model with Hot Deck Imputation achieves the highest accuracy at 96.43%, though it also has the longest processing time at 7.75 seconds. In contrast, the IBK model with Hot Deck Imputation offers a good balance of high accuracy (96.97%) and the fastest processing time (0.01 seconds). Overall, these results strongly suggest that Hot Deck Imputation significantly improves model performance across various algorithms compared to using datasets with missing values (No Imputation).