Abstract： Objective:To explore the application of metagenomic next-generation sequencing (mNGS) in patients with pulmonary infection after failure of empirical treatment.Methods:From September 2021 to November 2023, a total of 64 patients with pulmonary infection who failed to receive empirical treatment in the Ninth People′s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine were retrospectively analyzed, sputum and bronchoalveolar lavage fluid of patients were collected for traditional etiological detection and mNGS detection in alveolar lavage fluid, and the differences between traditional etiological detection methods and mNGS detection methods for pathogen detection in patients with pulmonary infection after failure of empirical treatment were compared.Results:In 64 patients with pulmonary infection after failure of empirical treatment, the positive rate of mNGS microbial detection in alveolar lavage fluid was higher than that of traditional etiological detection: 87.50%(56/64) vs. 57.81%(37/64), and the difference was statistically significant ( P<0.01). The most common microorganisms detected by mNGS were bacterial infections, the main bacteria were Pseudomonas aeruginosa, Klebsiella pneumoniae and Haemophilus paraininfluenzae. The detection rate of mNGS in mixed infection was higher than that of traditional etiological detection: 65.63%(42/64) vs. 15.63%(10/64), χ2 = 33.17, P<0.01. Drug resistance genes were detected by mNGS technique in 18 patients, and a total of 21 kinds of drug resistance genes were detected, 53.13%(34/64) of patients improved after antibiotic adjustment based on mNGS test results. Conclusions:mNGS technology can effectively improve the positive microbial detection rate of patients with pulmonary infection after failure of empirical treatment, and can assist in the evaluation of antimicrobial resistance genes and guide the adjustment of clinical antibiotics, so as to improve the therapeutic effect.