Overcrowding in emergency departments (EDs) is a global issue that negatively impacts patient care and operational efficiency. The triage process plays a crucial role in mitigating these challenges by determining the urgency of patients’ conditions. This study proposes a machine learning-based enhancement of the Emergency Severity Index (ESI), focusing on improving the accuracy and reliability of identifying low-urgency patients. We developed a predictive model using Decision Tree and Random Forest classifiers, trained on patient symptoms and vital signs. Our results indicate that the Decision Tree classifier achieved 100% accuracy, while the Random Forest classifier reached 94% accuracy. Additionally, we implemented a web-based dashboard to streamline the triage process, allowing real-time data input and decision-making support for triage officers. By leveraging machine learning, this approach significantly reduces the risk of misclassification in triage, improves patient throughput, and optimizes resource allocation. The study emphasizes the importance of integrating advanced predictive models into healthcare systems to address the growing issue of emergency department overcrowding. Further validation using real-time clinical data is suggested to enhance model robustness and applicability in diverse healthcare settings.