Abstract: With the Development of the Atmospheric Integrat ion Sounding System in China, more and more stations measure meteorological elements by the Automatic Weather Stations (AWS).However, the long-range data w ere obtained from manual observing.Analy sis and evaluation of differences betw een two kinds of data are impo rtant for estimating the impacts of changing inst ruments on measurements and for providing scientific evidences for the continual use of the data after changing instruments.Based on the parallel data of automatic sounding and manual observing in 2000, a comparat ive analysis is made on the validity and continuity of AWS data.The results indicate that generally speaking, departures between AWS and man-made observations are acceptable and their seasonal and regional dif ferences are negligible; owing to the higher sensitivity of AWS data, the departures are larger at sunrise and sunset, as well as at times of maximum and minimum temperature occurring; there is no obvious difference between AWS data and historical sequence.It can be concluded that automatic observat ion can be put to operational use, but needs parallel observing for a certain time, especially in Northwest China and mountain regions.