Pneumonia and respiratory tract infection could be induced by many types of pathogens, which is also related to a variety of diseases. The accurate classification of pneumonia is a key issue in the diagnosis and treatment. In this study, we studied the classification and diagnosis of pneumonia based on Raman spectroscopy of respiratory mucus. The Raman spectra of respiratory mucus samples from normal person and patients with common pneumonia and pneumonia complicated with plastic bronchitis were examined respectively. Then, the positions and intensities of the characteristic peak in Raman spectrum were analyzed to access the composition and molecular bonding changes corresponding to the glycosylation and fibrosis process of mucin e. The classification model of pneumonia was studied by combining principal component analysis and chemometrics methods. The experimental results show that the accuracy of the method proposed in this paper can reach 99.08% for the classification of pneumonia. According to the confusion matrix, the accuracy of the classification for common pneumonia and pneumonia complicated with plastic bronchitis is higher, reaching 100% and 97.4%, respectively. The classification method of pneumonia based on molecular spectroscopy also provides a reference for the application of Raman spectroscopy technology in the diagnosis of infectious diseases.