Partial least squares (PLS) method was applied to establish models with near infrared spectroscopy (NIRS) for quantitative analysis of polysaccharides,protein,cordycepic acid,adenosine content in cordyceps militaris mycelium. The origin spectra were used to establish the PLS models for quantitative analysis of adenosine and polysaccharides content in cordyceps militaris mycelium. The Savitzky-Golay smoothing spectra were applied to establish the PLS models for quantitative analysis of protein and cordycepic acid content. The values of root mean square error of cross-validation(RMSECV) of the most optimum PLS quantitative analysis models of polysaccharides,protein,cordycepic acid and adenosine were 0.0091,0.0222,0.0088 and 0.6520,the values of root mean square error of prediction(RMSEP) were 0.0079,0.0196,0.0087 and 0.5780. The results demonstrated that the quantitative analysis model for determining the four contents of cordyceps militaris mycelium possessed good stability and high precision. This method has great foreground in quantitative analysis of Chinese traditional medicine.