In order to understand of the influence of different calculation methods on the calculated daily mean temperature, an analysis is made of the difference in daily average temperature calculated from the observations between 24 and 4 times per day, as well as 4 and 3 times per day, by using the temperature obtained by manual observation data from 1961 to 2010 from 6 national base stations in Shaanxi. The homogeneity Penalized Maximal 〖WTBX〗F〖WTBZ〗 Test (PMFT) is carried out with the long temperature sequence, which consists of manual 4 and 24 times per day observation or 3 and 4 times per day times observation. The results indicate that the 24 times per day average temperature is higher than that from the 4 times per day calculation; the average difference is 013 ℃; and standard deviation is 039 ℃, with the biggest difference in autumn. The 4 times per day daily average temperature is 014 ℃ lower than that calculated by 3 times per day observation, and the standard deviation is 085 ℃, with the biggest difference value in summer. The monthly or annual mean temperature calculated from temperature observations with different number of observations varies, with the difference being about± 02 ℃. If the 24 times per day temperature is used, for a single station, the temperatures increasing rate may be 003 to 004 ℃/10a higher. Nevertheless, the calculation method (changing from 4 times to 24 times, or from 3 times to 4 times) will not change the inhomogeneity of a temperature sequence.