The health of moms is an important part of public health because it affects both their own health and the health of their children in the long run. Assessing the health risk of a mother can be made easier with the help of predictive analytics. This helps doctors find high-risk pregnancies and act on them quickly, which leads to better health results. In this work, we suggest a new way to look at the risk to a mother's health by using the Random Forest algorithm and adjusting the hyperparameters. Random Forest works by using random parts of the data and traits to build various decision trees. Each tree guesses the target label on its own, and the final prediction is made by voting (classification) or average (regression) the results, which makes for more accurate and reliable forecasts. The hyperparameters n_estimators and criterion play vital roles in determining the performance of the Random Forest model. To achieve optimal results, an extensive hyperparameter tuning process using cross-validation and grid search techniques is performed. This procedure identified the best combination of hyperparameters, leading to a robust model with improved generalization capabilities. The proposed model achieved an impressive accuracy of 90.15% in classifying maternal health risk, outperforming several benchmark models and demonstrating its potential as a reliable tool for healthcare practitioners.