Abstract: Dam deformation observation data can be regarded as non-stationary time series, and considering the influence factors of dam deformation, it can be decomposed into principal value function terms, periodic function terms and improved stationary time series. As the semi-empirical formula of time effect factors can not accurately fit the actual changes, a wavelet analysis is made to decompose the series into low frequency and high frequency signals: the low frequency signal represents time effect deformation trend while the high frequency signal represents the changes in water level and temperature. Then the deformation prediction ARMA model can be established by the time series analysis. The tendency of dam deformation can be predicted based on the water level and temperature observation data in the future. The actual calculated results show that the prediction reliability which is established by the time series analysis method is better than that by the regression analysis model; and the wavelet analysis method can be used as an effective method for dam deformation prediction.