Abstract: It is of significance to understand the relationship between thermal infrared brightness temperature and land temperature from observation stations, in order to know the observational accuracy of brightness temperature and the difference among temperatures of different meaning. In this study, NOAA thermal infrared BT and meteorological data from meteorological stations are compared and studied in different ways. The result shows: ① In the observation of BT from TIR, due to influence of short period variation such as weather or cloud, diurnal variation present in high frequency and sudden jump. However, the brightness curve fitted with maximum BT has better annual variation regularity. ② Land temperature of superficial layer shows regular daily and seasonal variation because of the influence of air temperature and solar radiation. ③ Annual variation of land temperature of deep layer is smooth and is correlated with seasonal variation. However, in comparison with air temperature, there is a hysteretic effect, and the deeper the layer is, the longer the lag time. ④ BT and air temperature, and 0.2 m land temperature, reveal a quite good correlation. The extreme value appears almost at the same time, corresponding to seasonal variation. It indicates that BT can reflect variation of land surface temperature, actually providing reliable evidence for using thermal infrared brightness temperature to extract abnormal information of earthquakes.