Based on new type, efficient and biological compatible near infrared surface enhanced Raman scattering (NIR used-SERS), the oxyhemoglobin of healthy people and patients with lung cancer were studied with NIR-SERS spectra. The study found that the NIR-SERS spectra have evident differences in certain areas, using principal component analysis (PCA) and independent variable T test statistical analysis method, NIR-SERS spectra in the principal component scores PC3, PC4 and PC9 has significant differences, and through the PCA score 2d and 3d scatterplot can differentiate healthy people and patients with lung cancer. The diagnostic specificity was 88% while the result of sensitivity was 80%, with total discriminant accuracy of 84%. Through the attribution analysis, significant differences both used NIR-SERS spectra is related to the vibration of oxygenated hemoglobin and vinyl pyrrole ring groups (CaCm). The study provides experimental basis for lung cancer SERS spectroscopy diagnosis, NIR-SERS technology is expected to become a kind of new clinical auxiliary diagnosis tool.