Abstract: Abstract: To evaluate Valencia orange mastication degree and establish a model for predicting the sensory mastication of Valencia orange quantitatively, nine kinds of Valencia orange, the number of each kind was 30, a total of 270 Valencia orange samples were collected. The research measured the content of crude fiber, sensory attributes, and textural properties of Valencia orange. Sensory evaluation was performed by a panel including eight trained people, and the average score of flesh attributes, residue, and mastication was recorded. Simulating the process of chewing the flesh, compression experiments, TPA tests, and shear tests were performed to analyze the textural properties. The averages and the standard deviations of the three tests were calculated. Statistically, differences were found for crude fiber content among all the cultivars of Valencia oranges. The crude fiber content of Frost and Campbell was significantly higher than others in the condition of significance levels (p<0.05). The results of the sensory evaluation showed that the variation coefficient of flesh attributes, residue, and mastication was 0.4 or higher. The other textural indicators were significantly different, except for flexibility and adhesion in texture trails. Simple correlation analysis was performed between sensory evaluation, crude fiber content, and texture property parameters using SPSS software. The results indicated that the mastication degree and texture properties showed a significant correlation. Compression resistance, elastic modulus, hardness, springiness, shear force, and shear work were selected as a texture index to build a model. Collinearity diagnostics and principal component analysis were performed to eliminate collinearity, which was caused by a quite high correlation between textural indices. According to the feature vector of principal component and scores of each texture property parameter, the best subgroup of principal component factors was selected to build the regression model. Then, with principal component factor scores as independent variables, and the standardized sensory scores as dependent variables, the regression analysis was used to establish a multiple linear regression equation. The determination coefficient (R2) of the model was equal to 0.73. The performances of the model was calibrated by validation of set data. Concerning the validation set, the determination coefficient (R2) and prediction standard deviation (S.E.P) were respectively equal to 0.85 and 0.17. The results presented demonstrated the greater potential of the texture properties for prediction of mastication degree of Valencia orange, which mean that the evaluation model of fruit mastication based on texture properties could accurately evaluate the mastication of Valencia orange. It was feasible by using textural properties to evaluate the fruit mastication of Valencia orange, instead of sensory evaluation. The research provided a reference for evaluating the mastication of Valencia orange.