Liver transplantation is widely acknowledged as the only effective treatment for end-stage liver disease, and infection is reportedly an important cause of postoperative death. Clinical use of metagenomic next-generation sequencing (mNGS) to diagnose postoperative infection and successfully guide drug therapy remains rare. This study included patients with infectious complications after liver transplantation from July 2019 to December 2020 and was divided into three groups: pneumonia, unknown fever, and others (including hepatic failure, kidney failure, cirrhosis after LT, and other postoperative complications that predispose to infection). The mNGS sequencing was used to detect microorganisms, and the results were compared with traditional culture. We found that mNGS yielded improved sensitivity over culture (85.19 vs. 22.22%; p<0.0001) but lower specificity (35.71 vs. 89.28%; p<0.0001). Among the 48 kinds of pathogens detected, the Torque teno virus 22 (15/122) was the most common, followed by Primate erythroparvovirus 1 (13/122). The top four bacteria included Klebsiella pneumoniae (n = 8), Enterococcus faecium (n = 5), Stenotrophomonas maltophilia (n = 4), and Escherichia coli (n = 4). Aspergillus fumigatus was the most common fungus. The bronchoalveolar lavage fluid (BALF) exhibited the highest proportion of positive findings among sample types, with viral, fungal, and bacterial mixed infection being the most common (n = 6, 19.35%). Besides, using mNGS for early diagnosis of infection after liver transplantation may effectively prolong patient survival. This is the first study to explore the application value of mNGS and its comparison with traditional culture in pneumonia and other infections in post-liver transplantation patients. The simultaneous application of these two methods suggested that the Torque teno virus 22, Klebsiella pneumoniae, and the Aspergillus fumigatus are the most common pathogens of viruses, bacteria, and fungi after LT, suggesting that these pathogens may be associated with postoperative pathogen infection and patient prognosis. The mNGS technique showed distinct advantages in detecting mixed, viral, and parasitic infections in this patient population. Further studies are warranted to systematically elucidate the dynamic evolution and molecular characteristics of infection after liver transplantation.