Objective To evaluate the application of bedside ultrasound in evaluating diaphragmatic function and pulmonary ventilation, and the outcome of withdrawal of mechanical ventilation in patients with respiratory failure in combination with respiratory mechanical parameters such as airway occlusion pressure (P0.1) and rapid shallow breathing index (RSBI). Methods A total of 45 patients who underwent mechanical ventilation in Department of Critical Care Medicine, Shanghai East Hospital from September 2018 to December 2019 were selected. All the patients met the indications for spontaneous breathing trail (SBT) before extubation. After SBT for 30 min, ultrasound examination was performed to evaluate diaphragmatic function and pulmonary ventilation before withdrawal of mechanical ventilation, while respiratory mechanical parameters were recorded. According to the presence of re‑intubation within 48 h, the patients were divided into two groups: a successful group and a failed group. The receiver operating characteristic (ROC) curve was plotted to analyze pulmonary ultrasound score, diaphragmatic mobility, diaphragmatic thickening fraction, and the accuracy of P0.1 and RSBI in predicting withdrawal. Results There were 25 out of 45 patients who successfully took off the machine. Patients in the failed group presented decreases in diaphragmatic mobility and diaphragmatic thickening fraction, and increases in pulmonary ultrasound score, P0.1 and RSBI, compared with the successful group (P<0.05). According to the ROC curve, the sensitivity of pulmonary ultrasound score, diaphragmatic mobility, diaphragmatic thickening fraction, P0.1 and RSBI in predicting successful withdrawal was 85%, 80%, 96%, 70% and 90%, respectively. The specificity of the above five indicators was 72%, 80%, 60%, 76% and 68%. The cut‑off values of pulmonary ultrasound score, diaphragmatic mobility, diaphragmatic thickening fraction, P0.1 and RSBI ROC curves were used as predictive values. The predicted failed extubation was scored as 1 point, and the predicted successful extubation was scored as 0 point. The total cut‑off values of the above five indicators were added. The area under the ROC curve was 0.909, with a sensitivity of 90% and a specificity of 76%. Conclusions The combined use of ultrasound during withdrawal of mechanical ventilation to evaluate pulmonary ventilation, diaphragmatic function and respiratory mechanical parameters can reduce the risk of re‑intubation in patients with respiratory failure, which provides an optimal scheme in guiding withdrawal of mechanical ventilation to improve the outcome.