【Objective】 This research was conducted to establish an easy and effective evaluation system for longan storage quality based on physicochemical indices. 【Method】 Multivariate statistical analyses, including factor analysis, regression analysis and hierarchical clustering, were used to evaluate and grade the fruit storabilities of 30 longan varieties, and the mathematical models were established. 【Result and conclusion】 The results showed that the comprehensive scores for storability of 30 varieties were obtained, including normal-temperature pericarp browning and mass loss rate, low-temperature pericarp browning and aril autolysis, which were all significantly related with their corresponding indices of storage performance in different periods. The mathematics models of storage performance of longan fruits at different temperatures were established, from which the effective predictive indices were screened and a physicochemical index database was built. It is concluded that there are 3 physicochemical indices including maturity, a^* value of exocarp and total soluble solids (TSS) content for normal temperature storability, and 3 indices,i.e. skin crevasse crack, total sugar content in the aril and flavor for low temperature storability. The 30 varieties are classified into 4 types based on their storabilities at normal and low temperatures.