In consideration of the fact that the interaction between different monitoring points is seldom considered in traditional displacement monitoring, the application of the spatial autoregressive model in the prediction of high slope deformation was examined based on the basic theory of spatial econometrics. Using a high slope as an example, the results of displacement prediction based on the monitoring data of exterior deformation were compared with those from the traditional autoregressive integrated moving average model. The results show that, with a significant spatial autocorrelation coefficient, the spatial autoregressive model is more accurate than the traditional model in prediction of slope deformation, and that, compared with the traditional model, the spatial autoregressive model has more concise parameters, considers the factors more comprehensively, and can perform displacement prediction at different monitoring points at the same time.