Abstract: Landslides are a common geological hazard that seriously threatens the safety of people’s lives and properties.Hence,the accurate prediction of landslide displacement is imperative to minimize landslide losses.In this study,a multivariate model is developed for landslide displacement prediction by combining a convolutional neural network (CNN) with long short-term memory (LSTM),conducting a principal components analysis to reduce dimensionality,and performing Bayesian optimization on the hyperparameters.Taking Baijiabao landslide as a specific case study,12 sets of monitoring data of Baijiabao landslide from 2017 to 2019 are selected as the basis to construct and compare a univariate CNN-LSTM model,multivariate LSTM model,multivariate CNN model,and multivariate CNN-LSTM model.The results show that the multivariate CNN-LSTM model greatly outperforms the other models in terms of the key performance assessment indicators,including the mean absolute error,root mean square error,mean absolute percentage error,and coefficient of determination of the prediction results.In addition,the fitting curve of values predicted by the multivariate CNN-LSTM model is the closest to that of the true values in the test set.The results of this study show that the multivariate CNN-LSTM model is superior to the other models in the displacement prediction of Baijiabao landslide.Therefore,the multivariate CNN-LSTM model is a novel and effective solution in landslide displacement prediction and a powerful tool for practical disaster prevention and mitigation work.