The action mechanism of external factors on the change of landslide displacement quantity is very complex, in order to establish an efficient and accurate landslide displacement quantity prediction model. This paper will analyze 35 possible landslide displacement factors from temperature, soil temperature, humidity, soil moisture and rainfall. First, the data was processed with missing values and outliers, data segmentation, principal component analysis-dimensionality reduction, to obtain the main influencing landslide displacement factor. Then compare the regression predicted performance of MARS, Random Forest, ANN and SVM. The final regression prediction is ideal for the Random Forest and SVM algorithms.