With the support of the Third Tibetan Plateau Experiment of Atmospheric Sciences, variations of the atmospheric boundary layer height of Nagqu area in summer(Jul-Aug 2014) were compared. Ceilometer derived atmospheric boundary layer height was estimated via gradient method of backscatter signal. The result shows that, when the cloud top height is determined as atmospheric boundary layer height, the upper limit of convection boundary layer height could reach 3800 m. And, for nocturnal stable boundary layer, the lower limit is only 40 m. Radiosonde derived atmospheric boundary layer height was obtained using gradient method of potential temperature and relative humidity. By analyzing the valid dataset derived from the simultanious measurements of ceilometer and radiosonde within the 40 days, the correlation coefficient of atmospheric boundary height obtained by gradient method of backscatter signal and potential temperature, is calculated as 0.78, while the standard deviation is 738.84 m. In addition, it has been found that signal-to-noise ratio of ceilometer detection is insufficient under clear sky condition by analysis of 8 sets of abnormal data. In short, results indicate great diurnal variation of atmospheric boundary layer height by comparing daily measured data, which could be attributed to that the high altitude of Nagqu district causes stronger solar radiation compared to plain area, and hence intense diurnal variation of temperature and convection in summer.