In　view　of　the　shortcomings　in　the　current　research　on　the　accuracy　of　landslide　disaster　early　warning　and　the　relationship　between　multiple　landslides,　this　paper　combines　the　collaborative　filtering　method　and　the　logical　regression　training　method　to　improve　on　the　previous　research　on　early　warning　only　for　a　single　landslide.　The　results　show　that　the　combination　of　the　collaborative　filtering　method　can　use　the　similarity　between　landslides　to　predict,　and　the　model　training　results　have　been　significantly　improved.　The　prediction　of　landslide　disasters　has　a　higher　recall　rate,　which　is　of　great　significance　to　the　prevention　and　control　of　landslide　disasters.　©　2023　IEEE.