Background: Ultrasound is a convenient tool to evaluate cardiac and diaphragm function. The ratio (E/Ea) of mitral Doppler inflow velocity to annular tissue Doppler wave velocity by transthoracic echocardiography (TTE) and diaphragmatic excursion (DE) by diaphragm ultrasound have been confirmed in predicting extubation outcomes independently, however their different roles in the weaning process have not been determined until now. Methods: We designed a cohort study to preform diaphragm ultrasound and TTE before and after the spontaneous breathing trial (SBT) in difficult-to-wean patients. Patients considered for enrollment should succeed on a SBT and have been extubated. They were followed up with the events of respiratory failure within 48 h, and divided into the respiratory failure and extubation success subgroups. Relevant risk factors predicting respiratory failure were analysed by a multivariate logistic regression model. Then, each subgroup was assessed with respect to re-intubation within 1 week, and divided into the re-intubation and non-intubation subgroups. Furthermore, relevant risk factors predicting re-intubation were also analysed in each subgroup. The area under the curve (AUC) and optimum cut-off value were identified by the receiver operating characteristic curve. Results: Among 60 patients, 29 cases developed respiratory failure within 48 h, and 14 cases were re-intubated or died within 1 week, respectively. Multivariate logistic regression analysis showed that E/Ea (average) after SBT [odds ratio (OR) 1.450, 95% confidence intervals (CI) 1.092-1.926, P = 0.01] and left ventricular ejection fraction were associated with respiratory failure. The AUC of E/Ea (average) after SBT was 0.789, and a cut-off value >= 12.5 showed the highest diagnostic accuracy with a sensitivity and specificity of 72.4% and 77.4%, respectively. Furthermore, in the respiratory failure subgroup only DE (average) after SBT was associated with re-intubation (OR 0.690, CI 0.499-0.953, P = 0.024). The AUC of DE (average) after SBT was 0.805, and a cut-off value <= 12.6 mm showed the highest diagnostic accuracy with a sensitivity and specificity of 80% and 68.4%, respectively. Conclusions: E/Ea (average) after SBT could help predict respiratory failure within 48 h. However, DE (average) after SBT could help predict re-intubation within 1 week in the respiratory failure subgroup.