Pregnancy complications are an essential challenge in obstetrics as they may jeopardize both maternal and neonatal outcomes if these conditions are not recognized and intervened in time. An early and precise prediction of pregnancy complications and timely intervention can decrease the risks associated with pregnancies. For early risk detection machine learning models are used to predict maternal health risks and also use the Genetic Algorithm (GA) approach to improve the predictive precision. Seven machine learning algorithms are optimized with GAs to determine their effectiveness in maternal health risk detection. In this study, the Decision Tree algorithm showed the highest accuracy at 94.16%, and the K-Nearest Neighbors (KNN) algorithm showed the lowest accuracy at 84.9%. This paper provides a clear methodological framework, focusing on implementing GAs in optimizing machine learning models. The findings bring out the promise of this integrated approach to improving the prediction of risk in maternal health and thus contributing to reduced maternal mortality.