Invasive and non-invasive mechanical ventilation (MV) have been combined as sequential MV in the treatment of respiratory failure. However, the effectiveness remains unclear. Here, we performed a randomized controlled study to assess the efficacy and safety of sequential MV in the treatment of tuberculosis with respiratory failure.Forty-four tuberculosis patients diagnosed with respiratory failure were randomly divided into sequential MV group (n = 24) and conventional MV group (n = 20). Initially, the patients in both groups received invasive positive pressure ventilation. When the patients' conditions were relieved, the ventilation modality in sequential MV group was switched to oronasal face mask continuous positive airway pressure until weaning.After treatment, the patients in sequential MV group had similar respiratory rate, heart rate, oxygenation index, alveolo-arterial oxygen partial pressure difference (A-aDO2), blood pH, PaCO2 to those in conventional MV group (all P value > 0.05). There was no significant difference in ventilation time and ICU stay between the two groups (P > 0.05), but sequential MV group significantly reduced the time of invasive ventilation (mean difference (MD): - 36.2 h, 95% confidence interval (CI) - 53.6, - 18.8 h, P < 0.001). Sequential MV group also reduced the incidence of ventilator-associated pneumonia (VAP; relative risk (RR): 0.44, 95% CI 0.24, 0.83, P = 0.006) and atelectasis (RR:0.49, 95% CI 0.24,1.00, P = 0.040).Sequential MV was effective in treating tuberculosis with respiratory failure. It showed advantages in reducing invasive ventilation time and ventilator-associated adverse events.Chinese Clinical Trial Registry ChiCTR2000032311, April 21st, 2020.