Abstract: In this paper, an early warning system of slope landslide for ion-type rare earth mine was established by selecting key factors causing slope landslide. Such system was set up in the target mine to collect monitoring data. The time and scale of the cracks in the site were also analyzed. When both data were combined, the threshold value of warning for single factor was obtained. Entropy weight theory was used to determine the weight of each influencing factor, and the multi-index comprehensive scoring value of Xikeng rare earth mine in Anyuan was calculated to obtain the comprehensive warning grade of rare earth mine slope landslide. At the same time, the single parameter warning was corrected to improve its accuracy as a result of false alarms caused by abnormal fluctuation of sensor data. It could be integrated into the early warning software algorithm to make the warning system intelligent. This is of great significance to mine safety production.