<p>Thrombosis presents significant healthcare challenges due to its complex nature. Recent advancements in data-driven mathematical and computational models of blood clot formation offer promising insights. The integration of machine learning (ML) and computational methods in thrombosis research is still in its early stages, but it could leverage the strengths of both approaches. This systematic review followed the PRISMA methodology to assess studies that (i) utilized computational models, (ii) modeled blood clot formation or thrombin generation through the coagulation cascade, and (iii) incorporated ML algorithms. We identified 11 eligible studies that focused on platelet signaling, outcome prediction, thrombin threshold prediction, shear rate prediction, and multiscale modeling. Artificial neural networks and support vector machines were the most commonly used ML models. The hybrid approach combining ML and computational models is still nascent but shows significant promise for advancing thrombosis research. These models offer valuable insights for improving thrombosis diagnosis, prognosis, and treatment, particularly in the context of personalized medicine for hemostatic disorders. The integration of ML with computational models holds great potential for improving thrombosis management, but further research is needed. Future work should focus on enhancing the physiological realism of these models, incorporating patient-specific data, and addressing challenges related to data standardization and clinical implementation. The field is in its early stages but shows promising growth potential and is well positioned to advance precision medicine approaches in thrombosis and hemostatic disorders. </p> Graphical abstract <p></p>