In the era of big data, the data scale of landslide monitoring could reach above terabyte level, the traditional database and data mining technology could no longer meet the requirements of intelligent monitoring and early warning. To obtain early warning information with high reliability and real time by applying big data theory, mechanisms, models and methods as well as machine learning methods are the inevitable trends in the future. This study aimed to realise a real time and precise mid-long prediction of landslide displacement, proposed two distributed landslide displacement prediction models: DLDP-GBTs (distributed landslide displacement prediction with Gradient Boosted Trees algorithm) and DLDP-RF (distributed landslide displacement prediction with Random algorithm); the cross-validation method was also adopted to evaluate and adjust parameters to reduce the root mean squared error of the model predicted results. In addition, this study proposed the rapid selection of features by using XGboost model in distributed situations can improve the Model training efficiency under distributed condition. By comparing different regression algorithms models, it was found that the DLDP-GBTs model based on the gradient optimisation decision tree was better than the other two models in terms of accuracy and real-time performance, which meets the requirements under the big data background.