Based on correlation and regression analysis on daily temperature data in Hefei,Anhui Province in 2003, the extremely significant correlations among median temperatures, minimal temperatures, maximal temperatures and average temperatures are found, and the goodness of fit test of regression model is more than 96.8%. According to the regression analysis on daily median temperatures and average temperatures for many years in Hefei, the regression equation y = 10026x-0.4683, which passes the α＝0.001 level testing is established. The same disipline can be found in other areas, which fully indicates a universal regression relationship between daily median temperatures and daily average temperatures. In addition，a new formula of aridity index calculation Ye＝0.282054＋0.552118E/r is setup based upon the correlation and regression analysis on the ratio of evaporation to rainfall and the aridity index of accumulated temperature. This method can be employed to calculate conveniently either the annual aridity index based on the conventional meteorological data of evaporation and rainfall or the seasonal and monthly aridity index during the period of temperature below 10℃.The calculated results have been tested well.