Hypertensive disorders in pregnancy are recognized as one of the major complications during gestation, posing serious risks to both the mother and the fetus. These disorders can result in stillbirths and preterm deliveries among otherwise normal pregnancies and are considered the third leading cause of maternal mortality worldwide. However, their exact etiology remains largely unknown. The main objective of this study was to identify the demographic factors influencing the incidence of hypertension in pregnant women using data mining algorithms. The study database included 4,818 records and 80 features, extracted from electronic health records registered in the Tehran University of Medical Sciences health centers through the SIB system of the Ministry of Health and Medical Education. The study followed the CRISP-DM methodology for implementation. Due to class imbalance in the dataset, modeling was performed in two ways: (1) using basic algorithms such as C5.4 decision tree, ID3, CHAID, and artificial neural networks; and (2) using ensemble methods that combined bagging and boosting with the aforementioned algorithms. According to the developed models, the most significant predictors of hypertension in pregnant women included negative Rh factor, maternal age, nutritional habits (consumption of fruits, salt, and type of oil), history of preeclampsia, smoking, marital status, and presence of other hypertensive risk factors. The results showed that the hybrid model combining C5.4 and CHAID decision trees achieved the highest accuracy (75%) in classifying hypertensive cases. The bagging ensemble with C5.4 and ID3 improved accuracy by 4.17%, while the bagging–neural network combination increased it by 30%. Other models employing bagging and boosting techniques did not show notable improvements.