This study explores the use of machine learning (ML) to screen for ovarian cancer (OC) by leveraging risk factors. The research aimed to develop a predictive model using six ML algorithms—XG-Boost, Random Forest (RF), J-48, Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Artificial Neural Network (ANN)—to identify high-risk OC cases. The dataset included 1516 suspicious OC cases from six clinical centers in Sari, Iran, between 2015 and 2019. After preprocessing, including removing duplicates, imputing missing data, and selecting relevant features, the best-performing model was identified as XG-Boost, which achieved an area under the receiver operating characteristic curve (AU-ROC) of 0.93. The model demonstrated high sensitivity, specificity, and accuracy, making it the most effective for OC prediction. Key risk factors identified included family history of cancer, menopausal age, history of chest X-ray, personal history of breast cancer, and postmenopausal hormone therapy. The model was also tested on external data from two clinical centers, showing good generalizability with AU-ROC values of 0.85 and 0.89 for the two settings. The study highlights the potential of ML in improving OC screening and prevention by identifying high-risk individuals based on risk factors. The results suggest that ML-based approaches can enhance the accuracy and efficiency of OC screening, contributing to better public health strategies. The study also acknowledges limitations, including the use of a retrospective dataset and the potential impact of missing data, and suggests future research should use larger, more diverse datasets and incorporate more factors for improved prediction. Overall, the study demonstrates the effectiveness of ML in predicting OC risk and supports the development of preventive strategies based on risk factors.This study explores the use of machine learning (ML) to screen for ovarian cancer (OC) by leveraging risk factors. The research aimed to develop a predictive model using six ML algorithms—XG-Boost, Random Forest (RF), J-48, Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Artificial Neural Network (ANN)—to identify high-risk OC cases. The dataset included 1516 suspicious OC cases from six clinical centers in Sari, Iran, between 2015 and 2019. After preprocessing, including removing duplicates, imputing missing data, and selecting relevant features, the best-performing model was identified as XG-Boost, which achieved an area under the receiver operating characteristic curve (AU-ROC) of 0.93. The model demonstrated high sensitivity, specificity, and accuracy, making it the most effective for OC prediction. Key risk factors identified included family history of cancer, menopausal age, history of chest X-ray, personal history of breast cancer, and postmenopausal hormone therapy. The model was also tested on external data from two clinical centers, showing good generalizability with AU-ROC values of 0.85 and 0.89 for the two settings. The study highlights the potential of ML in improving OC screening and prevention by identifying high-risk individuals based on risk factors. The results suggest that ML-based approaches can enhance the accuracy and efficiency of OC screening, contributing to better public health strategies. The study also acknowledges limitations, including the use of a retrospective dataset and the potential impact of missing data, and suggests future research should use larger, more diverse datasets and incorporate more factors for improved prediction. Overall, the study demonstrates the effectiveness of ML in predicting OC risk and supports the development of preventive strategies based on risk factors.