Acute respiratory distress syndrome (ARDS) remains an unsolved problem in the intensive care unit (ICU), which can be treated with venovenous extracorporeal membrane oxygenation (VV-ECMO). We summarized retrospectively collected data from an institutional experience with VV-ECMO in patients with severe acute respiratory failure and identified the clinical parameters associated with successful ECMO weaning. Among the 45 cases who received ECMO for pneumonia (n = 19), exacerbation of interstitial lung disease (n = 11), ARDS secondary to sepsis (n = 8), aspiration pneumonitis (n = 2), postoperative ARDS (n = 3), and others (n = 2), 21 (46.7%) were successfully weaned from ECMO. In a univariate analysis median platelet (PLT) count at ICU admission (162 vs. 97 × 10/L; p = 0.046) and pre-day 1 (118.5 vs. 62.5 × 10/L; p = 0.046) was higher in the ECMO-weaned group than those in the weaning failure group. Using a PLT level of 70 × 10/L, the odds ratio for successful ECMO weaning was 11.0 (95% confidence interval [CI] 1.34-87.16; p = 0.023) in the multivariate analysis. Bleeding complication rates were similar between the two groups. High PLT counts at ICU admission and the day immediately before initiating ECMO might play a key role in successful weaning of VV-ECMO for severe acute respiratory failure. Further studies should evaluate the proper target PLT level to enhance ECMO outcomes.