One quick and accurate method of distinguishing broiler chicken and native chicken was studied due to the disorder of chicken market. The discriminated model of chicken classification was built based on near infrared (NIR) spectroscopy technology and cluster analysis. The effects of different pretreatment methods on the prediction results were also discussed. The results showed that the discriminated model of chicken classification could be built based on NIR Spectroscopy technology and cluster analysis. The model could be used to distinguishing the broiler chicken and native chicken. Meanwhile, the rate of the data after the second derivative and vector normalization model prediction was 100%.