Over the past decade, there has been growing interest in extracorporeal membrane oxygenation (ECMO) as a rescue therapy for patients with severe acute respiratory distress syndrome (ARDS) and cardiogenic shock. Although survival of ECMO-treated patients has improved recently, the incidence of ECMO-related complications such as bleeding and nosocomial infections remains unacceptably high. In addition, patients often experience long-term physiological and psychological sequelae. Hence, identifying patients who will most likely benefit from ECMO is crucial. Because the technique exposes patients to complications and is associated with high costs and resource utilisation, prediction models have been developed to assist clinicians in identifying patients that would most likely survive after ECMO treatment. In addition, these prediction models enable comparison of risk-adjusted outcomes, both over time and between centres. Our review explores the latest predictive survival models developed for ECMO-treated severe cardiogenic shock and ARDS patients.