Currently, little in-depth evidence is known about the application of extracorporeal membrane oxygenation (ECMO) therapy in coronavirus disease 2019 (COVID-19) patients. This retrospective multicenter cohort study included patients with COVID-19 at 7 designated hospitals in Wuhan, China. The patients were followed up until June 30, 2020. Univariate and multivariate logistic regression analyses were performed to identify the risk factors associated with unsuccessful ECMO weaning. Propensity score matching was used to match patients who received veno-venous ECMO with those who received invasive mechanical ventilation (IMV)-only therapy. Of 88 patients receiving ECMO therapy, 27 and 61 patients were and were not successfully weaned from ECMO, respectively. Additionally, 15, 15, and 65 patients were further weaned from IMV, discharged from hospital, or died during hospitalization, respectively. In the multivariate logistic regression analysis, a lymphocyte count <= 0.5x10(9)/L and D-dimer concentration >4x the upper limit of normal level at ICU admission, a peak PaCO2 >60 mmHg at 24 h before ECMO initiation, and no tracheotomy performed during the ICU stay were independently associated with lower odds of ECMO weaning. In the propensity score-matched analysis, a mixed-effect Cox model detected a lower hazard ratio for 120-day all-cause mortality after ICU admission during hospitalization in the ECMO group. The presence of lymphocytopenia, higher D-dimer concentrations at ICU admission and hypercapnia before ECMO initiation could help to identify patients with a poor prognosis. Tracheotomy could facilitate weaning from ECMO. ECMO relative to IMV-only therapy was associated with improved outcomes in critically ill COVID-19 patients.