Background <p>Respiratory sequelae, induced by lung injury, reduced muscle strength, and nutritional disturbance, are common in hospitalized patients with coronavirus disease 2019 (COVID-19). Therefore, optimal treatment is essential for reducing the mortality in severe forms of the disease and critically ill patients. Pulmonary rehabilitation (PR) has been used in many chronic respiratory diseases, but the role of early PR in severe and critically ill COVID-19 patients remains to be fully understood.</p> Methods <p>Hospitalized severe to critically ill COVID-19 patients were recruited from Beijing Chaoyang Hospital between December 1, 2022, and June 30, 2023. In all, we recruited 272 patients, with 39 in the PR group and 233 in the control group. The PR intervention consisted of the prone position, airway clearance therapy (ACT), and resistance respiratory training (RRT). The primary outcome was the composite disease progression outcome rate, defined as death or intensive care unit (ICU) admission. Adverse events (AEs) and serious adverse events (SAEs) were recorded in the PR group. Inverse probability of treatment weighting (IPTW) and propensity score matching (PSM) was used to balance confounding bias, generating weighting cohort and matched cohort.</p> Results <p>The rate of the primary outcome was lower in the PR group (28.2% [11/39] in the PR group vs. 48.9% [114/233] in the control group). Significant differences were observed in both the original and weighting cohorts. Subgroup analyses showed that receiving ≥ 2 types of PR, receiving RRT, length from admission to intervention ≤ 4 days, and baseline P/F ≤ 150 mmHg were associated with lower rates of progression. Total rates of 2.6% (1/39) for AEs and 10.26% (4/39) for SAEs were reported.</p> Conclusions <p>Early pulmonary rehabilitation may prevent disease progression and reduce mortality in patients with severe COVID-19. These findings may be helpful for formulating an optimal rehabilitation strategy.</p>