Preeclampsia is one of the most common causes of maternal and perinatal morbidity and mortality. The disease is characterized by the development of hypertension after 20 weeks of gestation and can trigger a series of severe complications, such as eclampsia, kidney injury, liver damage, multiorgan dysfunction, and fetal growth restriction. In more severe cases, the condition may lead to preterm birth, significantly increasing the risk of neonatal complications. Therefore, early detection and proper management of the disease are essential to minimize adverse outcomes for both the mother and the babys. Traditional methods, such as blood pressure measurement, proteinuria analysis, and risk calculators based on regression methods, have limitations in predicting the disease before symptom onset, highlighting the need for more accurate approaches. In this context, machine learning emerges as a promising alternative to enhance preeclampsia detection. This approach enables the analysis of large volumes of clinical data and biomarkers, improving predictive capabilities. This study aims to map the application of machine learning in preeclampsia prediction, evaluating the advantages of these models over traditional methods and the challenges of their clinical implementation. A comprehensive scoping review will analyze the main models used, such as decision trees, Support Vector Machine, and Random Forest, as well as the most relevant predictive variables and performance metrics, including accuracy and the area under the ROC curve (AUC). Additionally, this study will contribute to identifying gaps in the literature and guiding future research to refine predictive models and improve their precision indicators.