In-hospital cardiac arrest (IHCA) is a critical medical issue threatening the survival and prognosis of hospitalized patients, characterized by high incidence, high mortality and poor prognosis. Early warning and intervention for IHCA are urgently needed. The early warning score (EWS) is developed as a point-of-care warning tool for early identification and intervention of hospitalized patients with deteriorating condition. In recent years, EWS has become one of the important methods for early warning of IHCA, especially EWS based on machine learning (ML) has shown great potential. This review mainly focuses on the traditional EWS and ML-based EWS, discusses the research status of EWS worldwide, and focuses on the research progress of EWS in early warning of IHCA.