Objective To evaluate the value of microbiological rapid onsite evaluation (M-ROSE) combined with bronchoscopy in the rapid diagnosis of pathogen in severe pneumonia and the optimization of treatment. Methods A total of 188 patients with severe pneumonia from Dali People’s Hospital from January 2024 to December 2025 were included. They were divided into the M-ROSE group (n＝88) and the control group (n＝100) based on whether they received M-ROSE treatment. The control group received routine empirical anti-infection treatment and underwent bronchoalveolar lavage fluid (BALF) culture as well as metagenomic next-generation sequencing (mNGS) testing. In the M-ROSE group, after obtaining BALF through bronchoscopy, the samples were immediately processed for staining and interpretation, and early anti-infection treatment was guided accordingly. The clinical outcomes, antibiotic usage, costs, and inflammatory indicators (white blood cell count (WBC), C-reactive protein (CRP), procalcitonin (PCT)) of the two groups were compared, and the diagnostic efficacy of M-ROSE was evaluated. Results There was no statistically significant difference in the baseline data between the two groups (all P＞0.05). The 28-day mortality rate in the M-ROSE group was lower than that in the control group (18.2% vs 32.0%, P＝0.034), and the length of hospital stay, the number of days of antibiotic use, the proportion of combined medication, and the hospitalization cost were all reduced (all P＜0.05). On the 3rd and 7th days after treatment, the levels of WBC, CRP and PCT in the M-ROSE group were all lower than those in the control group (all P＜0.001). The diagnostic sensitivity of M-ROSE was 85.7%, the specificity was 92.3%, and the Kappa value was 0.75. Conclusion The combination of M-ROSE and bronchoscopy can quickly provide pathogenic clues, optimize the early anti-infection treatment for severe pneumonia, improve prognosis and reduce medical burden.