Abstract: Taste, the combination of various taste-producing components in tea, is one of the most significant elements for tea quality evaluation. Green tea, a non-fermented tea, is the most produced and consumed tea in China. The enzymes in the fresh leaves for making green tea are basically passivated after high-temperature killing, and the original quality components are maximum retained in the fresh leaves. Different tea cultivars have various taste for the different types, contents, and composition ratios of biochemical components in fresh leaves. In recent years, in order to meet the development needs of the Chinese tea industry and changes in the industrial structure, numerous high-quality tea cultivars can be cultivated to produce green tea. However, the diversification of tea cultivars has brought great difficulties to identify or evaluate taste quality of green tea. A sensory evaluation method for tea taste quality can be susceptible to subjective factors of tea evaluation personnel, and external environmental factors, indicating lacking some objectivity of its evaluation to a certain extent. In this study, a systematic comparative investigation was conducted to evaluate the content of the main taste components and Dot value (concentration of taste components in tea soup / taste components threshold) of green tea made from 25 tea cultivars or strains, and a principal component analysis method was used to establish prediction models for different tea cultivars or strains, and taste quality of green tea. The results demonstrated that caffeine and catechins were the dominate elements for the difference in the intensity of bitterness and astringency. While Epigallocatechin gallate (EGCG) and caffeine were the main contributor to bitterness, EGCG was also a main astringency contributor. Amino acid was the prominent factor for the difference in umami and sweetness of green tea from different tea cultivars or strains, where glutamic acid was the predominant contributor to umami. Principal component analysis showed that the cumulative variance contribution rate of 5 principal components were 79.018%. The top four principal components can be considered as Astringent and bitterness, refreshing factor, bitterness factor and umami factor. The taste comprehensive evaluation model was constructed based on the linear regression equation and contribution rate of the previous five principal components, to evaluate the characteristics of green tea. The predictive model evaluation results were similar to that of the sensory evaluation, indicating an extremely significant correlation (P<0.01) and the correlation coefficient index was 0.791. The scores in the prediction model can be utilized to fully distinguish the taste characteristics, such as bitterness, astringent, refreshing, and umami of the tested samples. Therefore, the prediction model can be used to accurately distinguish and evaluate the difference in taste quality and bitterness, astringent, refreshing, umami and other taste characteristics of the tested samples. The finding can provide a theoretical basis for the popularization and application of various tea cultivars or strains, and new ideas and facile methods for the scientific evaluation of the taste quality of green tea from different tea cultivars or strains.