Background: Whether Lung ultrasound (LUS) can be used for pathogenic diagnosis is still controversial. This was conducted to test the accuracy and reliability of ultrasound in the diagnosis of pneumonia and to clarify whether ultrasound has diagnostic value for the etiology. Methods: A total of 135 neonatal pneumonia patients with an identified pathogen and 50 newborns with normal lungs in the newborn intensive care unit of 10 tertiary hospitals in China were enrolled. The study ran from November 2020 to December 2021. The infants were divided into various groups according to pathogens, the time of infection, the gestational age, the severity of the disease. The distribution of pleural line abnormalities, pulmonary edema, and pulmonary consolidation, as well as the incidence of air bronchogram and pleural effusion based on LUS, were compared between the above groups and between the pneumonia and healthy control groups. Results: There were significant differences in pulmonary consolidation. The sensitivity and specificity of the diagnosis of severe pneumonia based on the extent of pulmonary consolidation were 83.3% and 85.2%, respectively. The area under the receiver operating characteristic curve for the identification of mild or severe pneumonia based on the distribution of pulmonary consolidation was 0.776. Conclusion: Lung ultrasound has good performance in differentiating the severity of neonatal pneumonia, but cannot be used for pathogenic diagnosis.