A new method for recognition of raw material cultivar of manufactured tea with Fisher discriminant classification using near infrared spectra. In this study, spectra of samples with different raw material cultivar(Longjing43# and other cultivar) were collected. 20 principal components were obtained by PCA. 8 Principal components by step wise was used to establish the Fisher function for discriminant classification to recognize the raw material cultivar of manufactured tea. The result showed that the function performed well in recognition of raw material cultivar of manufactured tea. The accuracy for recognition was 96.8% in calibration set. 93.5% was obtained for unknown samples in test set. The result proved that it was feasible to recognize the raw material cultivar with combined analysis of PCA and Fisher discriminant classification using near infrared spectra.