Different people exhibit different allergic reactions to various types of pollen. Therefore, this study presents a method to rapidly detect and classify the pollen particles in air. Herein, 465 Raman spectroscopic data associated with 42 pollen samples were obtained using a Raman spectrometer by considering common pollen as the research object. Subsequently, they were categorized as interfamily and intergeneric pollen according to their biological classification, and then classified and predicted. After the obtained spectral data were preprocessed, principal component analysis was used for extracting the spectral characteristic information, and a support vector machine recognition model was established. The prediction accuracy of interfamily pollen is 97.75%, and the pollen prediction accuracy of the genus Rosaceae is 90.47%, which indicates that Raman spectroscopy can be used to classify and identify pollen in a feasible manner.