Objective：This study aims to assess the utility of rapid on-site evaluation（ROSE） for transbronchial lung biopsy（TBLB）and endobronchial ultrasound-guided transbronchial needle aspiration（EBUS-TBNA） in cytological diagnosis of lung malignant tumor. Methods：Total 174 cases performed TBLB and EBUS-TBNA procedures prospectively collected between January 2018 to August 2019 at the Affiliated Nanjing Hospital of Nanjing Medical University. They were randomly divided into the two groups according to with or without ROSE. There were 106 cases in ROSE group，while there were 68 cases in non-ROSE group. Histological findings was the gold standard. The accuracy of TBLB or EBUS-TBNA with and without ROSE was compared. The sensitivity of the two methods in cytological diagnosis of lung malignant tumor was analyzed，and the differences between the two methods in cytological diagnosis of lung malignant tumor was compared. Results：Overall accuracy in ROSE group and non-ROSE group was 88.7%和77.9%，respectively. Sensitivity to detect malignant lesions in ROSE group and non-ROSE group was 90.1%和71.4%，respectively. The diagnostic accuracy rate and sensitivity rate in ROSE group were higher than those in non-ROSE group（P＜0.05），and the rate of secondary inspection in ROSE group was lower than that in non-ROSE group（P＜0.05）. The time of inspection in ROSE group was higher than that in non-ROSE group，but there were no siginificant differences between two groups. Specimen satisfaction（96.2%） of ROSE group was higher than that of non-ROSE group（86.8%），and there were significant differences between two groups（P＜0.05）. Conclusion：The combination of ROSE can improve the satisfaction of TBLB and EBUS-TBNA in the sampling process，improve the diagnostic accuracy rate of lung lesions，especially the sensitivity of cytological diagnosis for malignant tumor，which has certain application value in the diagnosis of lung tumors.