Lower Limb Varicose Veins (LLVV) are a common problem among healthcare providers. Once it occurs, it will seriously affect the productivity and the quality of life of the affected person. Therefore, early prediction and preventing of Lower Limb Varicose Veins can improve patient prognosis. This study constructed different machine learning models to explore their efficiency in predicting LLVV. Five prediction models were applied to the study, including a Support Vector Machine (SVM) model, XGBoost model, decision tree (DT) model, Random Forests (RF) model, and Logistic Regression (LR) model. Afterward, the performance of the obtained prediction models was evaluated by Cross-validation, ROC-AUC (area under the curve), Precision-Recall AUC. The results showed that XGBoost outperformed the other models, achieving the highest accuracy ( \(0.85 \pm 0.03\) ), ROC-AUC (0.98), and Precision-Recall AUC (0.97), demonstrating its strong predictive power. While models like Logistic Regression, Support Vector Machine, and Random Forest also performed well with AUC values above 0.9, XGBoost consistently delivered the best results. Moreover, our study was conducted on a limited set of demographic and clinical variables without considering other possible determinants of LLVV such as lifestyle factors or genetic susceptibility. Thus, the model still needs external verification research before its clinical application.