Although extracorporeal membrane oxygenation (ECMO) has emerged as a pivotal temporary extracorporeal life support in critically ill patients, it is associated with a high risk of mortality. Establishing a risk prediction model for ECMO-related mortality is crucial for early identification and timely intervention to improve survival. This study aimed to develop and validate a machine learning (ML) model to predict the risk of ECMO-related mortality among patients admitted to intensive care units (ICUs). Retrospective cohort study. In the ICUs of the three hospitals in Changsha and Changde, two cities of Hunan Province, China. Patients aged ≥ 18 years old undergoing ECMO who were admitted to the ICUs, and the duration of ECMO was ≥ 24 h. The study was conducted among 204 patients undergoing ECMO from January 01, 2020, to December 31, 2023. The patients were randomly divided into a training set (n = 142) and a validation set (n = 62) at a ratio of 7:3. The least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression were used to screen for factors associated with mortality. Three ML classification models and the Shapley Additive exPlanations (SHAP) interpretation were used to identify the optimal model and assess its diagnostic performance. The final model identified the following factors associated with ECMO-related mortality: body mass index (BMI), APACHE II, history of smoking, white blood cell count (WBC), activated partial thromboplastin time (APTT), red blood cell distribution width (RDW), and platelet (Plt). The Gaussian Naive Bayes (GNB) model was the optimal model, with an average area under the ROC curve (AUC) of 0.906 (0.893–0.926) in the training set and 0.888 (0.75–1.00) in the validation set. This study presents an internally validated, explainable machine learning model for predicting ECMO mortality, developed using a specific Chinese cohort. The model demonstrated strong performance within our study population and identified key factors associated with ECMO-related mortality.