Abstract: The daily mean temperature is the basis for the statistic analysis of monthly, annual and multi-year average temperature, and is also an important indicator for studying the characteristics of long- term temperature change. Understanding the differences between the daily mean temperature values obtained by different observation times or calculation methods is important for the study of climatology and climate change. Based on the analysis of minute air temperature data of Harbin station, the mean value of 24 times timing air temperature could represent daily mean temperature. Using the 24 h data of 8 national reference climatological stations, we analyzed the effect of observation times and calculation method on the evaluation of mean temperature in long-time scale. The results showed that: (1) the annual average temperature obtained by three times weighted value (T3) and the maximum and minimum average (Tmn) were all on the high side, and the annual average temperature obtained by four times average (T4) was closed to 24 times average (T24); (2) on seasonal average, T3 was higher in spring and summer, meanwhile lower in winter; Tmn was higher in summer, autumn and winter; compared with T24, T3 of the whole province was generally higher from March to October; Tmn was on the high side except May; T4 was higher from April to July; T3 and T4 were higher in the latter stage of the first half of the year, and lower in the beginning and the end of the year in each station; (3) Tmn was high throughout the year in southwest stations and in the beginning of the year and the second half of the year in northern stations; (4) T3 and Tmn were higher in warmer years.