Preterm birth (delivery before 37 weeks) affects about 11% of live births worldwide and is a leading cause of neonatal mortality and long-term developmental issues. While clinical interventions like cervical cerclage and progesterone help high-risk mothers, most preterm births occur without clear warning signs, making early prediction difficult. This study evaluates artificial intelligence (AI) and machine learning (ML) methods to improve preterm birth prediction. Using a retrospective case-control design with U.S. nationality data (n \(\approx \) 119,000), several ML models (logistic regression, random forest, XGBoost, support vector machine, neural network) classified pregnancies as preterm or term. Performance was measured by accuracy, precision, recall, F1 score, and AUC-ROC. The support vector machine (SVM) achieved the highest accuracy (approximately 84%) and AUC-ROC (approximately 73%) on an independent test set. Key predictors included maternal gestational hypertension, prior preterm delivery, and tobacco exposure, aligning with known risk factors. These results show AI models can improve early identification of at-risk pregnancies, aiding targeted interventions to enhance neonatal outcomes.