[Objective] The objective of this study is to explore the effect of the comprehensive application of the third theory of quantification and BP neural network in the landslide, in order to provide a new method for the prediction of landslide volume.[Methods] The influence factors of landslide volume and its coupling strength were analyzed by the third theory of quantitatification. Based on the analysis results, the secondary factors and strong coupling degree samples were removed, and then the BP neural network prediction models of 3 different kinds of landslide volume was built according to different stages of the elimination.[Results] The main influencing factors of landslide volume were slope angle, slope, vegetation coverage rate and slope high, while the secondary influence factors were the dip angle, elevation and slope rock orientation. And in different samples, the degree of coupling between the volume influencing factors was difference.[Conclusion] The prediction method used in the present study is feasible, and the prediction accuracy can be improved by eliminating the secondary factors and the strong coupling degree samples.