Abstract: 【Objective】 To promote the utilization of sand pear(Pyrus prifolia Nakai)germplasm of Guangxi landraces,twelve characters were analyzed and evaluated from 49 sand pear landraces maintained at the National Sand Pear Germplasm Repository.【Method】 The coefficient of variations and Shannon-Weaver indexes of 12 fruit quality characters were analyzed. Q cluster and R cluster analyses were carried out to classify germplasm resources and traits and principal component analysis was used to evaluate fruit quality.【Result】 The coefficients of variation of fruit quality traits in Guangxi landraces varied between 9.25% to 52.57% and their Shannon-Weaver indexes ranged from 1.827 to 2.037. The average values of single fruit weight,fruit diameter,fruit length and fruit shape index were 223.2 g,7.19 cm,7.18 cm and 1.00,respectively. The corresponding coefficient of variations for these were 33.20%,16.9%,11.26% and 12.24% and the Shannon Weaver diversity indexes were 1.851,1.965,1.907 and 1.949,respectively. The average values of flesh hardness,soluble solids,soluble sugar,titratable acid,Vitamin C and stone cell content were 12.00 kg/cm²,10.49%,5.63%,0.31%,1.75 mg/100 g and 0.90%,respectively. The coefficient of variations for such were 14.40%,9.25%,15.18%,52.57%,18.25% and 35.18% and the Shannon Weaver diversity indexes were 1.957,2.037,1.935,1.910,1.827 and 1.933,respectively. The soluble solids had the largest Shannon Weaver index(2.037),whereas the titratable acid content had the largest coefficient of variation(52.57%)and the richest diversity. The average values of solid acid ratio and sugar acid ratio were 42.73 and 22.65 with coefficients of variation of 45.50% and 44.76% and Shannon Weaver diversity indexes of 1.956 and 1.848. The 12 fruit quality traits were divided into 4 groups by R cluster analysis. The correlation analysis and R-type cluster analysis revealed that fruit quality displayed logical correlation between some traits and significant(P<0.05)or highly significant(P<0.01)correlations were observed among most traits. Four principal components were extracted from our principal component analysis and it covered 80.424% of the original data. A comprehensive evaluation model was established by principal component analysis and 9 cultivars were selected with a score above 1.000.The 49 sand pear landraces were divided into three groups by Q-clustering,corresponding to good,medium and poor quality.【Conclusion】 Abundant genetic diversity was observed in Guangxi landraces of sand pear through analyses of fruit quality traits. The results from the evaluation of sand pear fruit quality by principal component analysis in Guangxi landraces proved to be reliable and the evaluation model developed provides a theoretical reference for future comprehensive evaluations of fruit quality in sand pear landraces.