Blood is a crucial body fluid which contains erythrocytes and leukocytes and platelets, the number and status of which directly indicate the physiological state of the body. The first response to the infection is mediated by the number of leukocytes in the blood. The number and type of immune cells change vary in the disease state and identification of the type of immune cell provides information about the healthy state of body. Determining the identities of cells of the immune system usually involves destructive fixation and chemical staining, or labeling with fluorescently labeled antibodies. Raman microscopy is ideal for live cell studies or real-time diagnosis of disease, because it does not require the use of labels that may harm cells. It has potential to be carried out in vivo conditions. Raman spectroscopy has been used to investigate the differences between the leukocytes subtypes. The complex chemical composition of cells leads to complex Raman spectra, it is difficult to distinguish the categary of five subtypes white blood cells. We propose a partition principal component analysis (PPCA) method based on Raman spectroscopy using wavelet anlysis at the single cell level to separate Raman spectra of five subtypes of leukocytes, which are respectively lymphocytes, nuetrophils, monocytes, eosinophils and basophils. We achieved the identification and differentiation of five subtypes with a minimum discrimination efficiency of 85%. Systematic studies of five leukocyte subtypes have important guiding significance for the study of various leukocyte-associated cancers.