Abstract In patients intubated for hypoxemic acute respiratory failure (ARF) related to novel coronavirus disease (COVID-19), we retrospectively compared two weaning strategies, early extubation with immediate non-invasive ventilation (NIV) versus standard weaning encompassing spontaneous breathing trial (SBT), with respect to IMV duration (primary endpoint), extubation failures and reintubations, rate of tracheostomy, intensive care unit (ICU) length of stay and mortality (additional endpoints). All COVID-19 adult patients, intubated for hypoxemic ARF and subsequently extubated, were enrolled. Patients were included in two groups, early extubation followed by immediate NIV application, and conventionally weaning after passing SBT. 121 patients were enrolled and analyzed, 66 early extubated and 55 conventionally weaned after passing an SBT. IMV duration was 9 6–11 days in early extubated patients versus 11 6–15 days in standard weaning group ( p  = 0.034). Extubation failures 12 (18.2%) vs. 25 (45.5%), p  = 0.002 and reintubations 12 (18.2%) vs. 22 (40.0%) p  = 0.009 were fewer in early extubation compared to the standard weaning groups, respectively. Rate of tracheostomy, ICU mortality, and ICU length of stay were no different between groups. Compared to standard weaning, early extubation followed by immediate NIV shortened IMV duration and reduced the rate of extubation failure and reintubation.