To study the spectrum processing methods for Raman microscopy on single living cell and develop the pre-process techniques for Raman spectrum of single living cell to enhance the signal to noise ratio, sensitivity, and decrease the fluorescence influence, wiping off the cosmic rays was used to improve the spectrum. The spectra classification, spectra average and filtration were applied to enhance signal to noise ratio. The fluorescence was depressed for quantity analysis or utilized for analysis by comparing the background and the spectra. Results show that (1) comparing the spectra with short exposure time and more scans can wipe off the cosmic rays effectively. (2) the spectra classification, spectra average and filtration can improve the quality of spectra and can show some weak and sensitive bands. (3)sometimes the fluorescence has useful information. It is concluded that the proposed techniques for Raman spectrum of single cell in this paper can show the sensitive and weak intensity peaks and reflect the information of molecules structures very well.