In order to estimate the quality of the fresh tea leaves by near infrared spectrum (NIRS) technique, the correlative model equation between near infrared spectrum of the fresh tea leaves and their corresponding moisture, crude fiber and total nitrogen amounts respectively through cross-validation and partial least squares (PLS) method were established in this paper. When the dimension was 7, the three indicators of R value were 0.92, 0.86 and 0.90, root mean square error of cross-validation (RMSECV) were 0.769, 0.332 and 0.742, respectively. A quality coefficient equation was suggested depending on the contents of moisture, crude fiber and entire nitrogen of fresh tea leaves. According to the quality coefficient equation of fresh tea leaf raw material, the quality coefficient of fresh tea leaves was obtained. The larger quality coefficient is, the higher quality of fresh leaf is. Thus, we developed a new method for the evaluation for the quality of fresh tea leaves by using NIRS was developed.