Abstract: In order to comprehensively evaluate the difference in fruit quality among 16 different feijoa lines, this study determined the external and internal quality indicators and antioxidant activities of 16 feijoa lines. Variability analysis, correlation analysis and principal component analysis were used to identify core indicators. The weights of these indicators were determined by analytic hierarchy process, and a comprehensive evaluation model for fruit quality was established by grey relational analysis. The accuracy of the model was verified by sensory evaluation. The results indicated a significant variation in fruit quality across feijoa lines. Four core fruit quality indexes were selected, including single fruit mass (X1), juice yield (X2), sugar/acid ratio (X3), and total phenol content (X4). The comprehensive evaluation model was described as follows: weighted grey relational grade = 0.381 0X1 + 0.298 2X2 + 0.210 8X3 + 0.110 0X4. The model results were correlated well with those of sensory evaluation (R2 = 0.869), indicating high accuracy of the model. Two excellent lines, Y4 and Y1, were selected for promotion. The findings provide a reference for the establishment of a comprehensive evaluation system for feijoa quality and offer scientific guidance for feijoa breeding.

Key words: feijoa; fruit quality; analytic hierarchy process; grey relational analysis; comprehensive evaluation