Abstract: The R/S analysis (rescaled range analysis) is one of efficient methods to characterize nonlinear time series. As the time series of monitoring surface displacement is not only deterministic but also random, the R/S analysis is used to analyze the long-term displacement time series of a landslide in the Three Gorges Reservoir area. The Hurst index of the accumulated displacement-time series and the incremental displacement-time series of the landslide are calculated by using the R/S analysis. The incremental displacement-time series is a more efficient fractal structure, and its Hurst index can depict quantitatively the deformation trend of the landslide. The concept of long range correlated critical time is introduced to analyze the non-periodic cycle length of the landslide displacement. The research indicates that the Hurst index and the non-periodic cycle length at different places of the landslide are different. Based on the physical mechanism of landslide deformation evolution, a conceptual model to determine the accelerating deformation phase according to the Hurst index of the incremental displacement-time series is proposed, and it provides a new method to improve the prediction of landslide failure time.