Abstract: In this paper, a systematic recognition method is applied to the analysis of crustal deformation data.After recognition and study for the data fron Daiyuan, Dahuichang and Noukouyu stations, it is considered that the self-regression model with controlling term (called CAR (n) model for short) can cbjectively and reasonably reflect the crustal deformation observation data which overlap the temperature and rainfall influences.If the model is used in data estimation, disturbance removement and the recognition of seismic anomalies, the result will be relatively ideal.