Abstract: Due to the complex structure of rock and soil mass and various failure mechanisms,landslide early warning has always been a global problem of very challenging. At 5:53 a.m. on 17 February 2019 in Longjing Village,Maling Town,Xingyi City,Guizhou Province,a landslide geological disaster occurred. About 960,000 cubic meters of mountain had a bedding sliding,which posed a serious threat to the road and residential area located at the foot of the front slope. In this paper,we discuss the characteristics and successful early warning process of this Xingyi landslide in Guizhou Province,and analyze the key factors of successful early warning. On the basis of geological investigations on the landslide site,this paper comprehensively use satellite remote sensing,UAV aerial photography,lidar,surface displacement monitoring and other technical means to reveal the historical deformation and pre-sliding deformation characteristics of the landslide. Finally,we analyze the mechanism of the landslide. After preliminary analysis,we conclude that Xingyi landslide belongs to a typical bedding rock landslide with weak intercalation. The slope in the source area of the landslide was an unstable slope formed after the first sliding in 2014. It had a combination of unfavorable slope structure and weak intercalation. Under the action of long-term gravity and groundwater,it eventually evolved into a landslide geological disaster. In the process of landslide evolution,the strength of the weak intercalation decreased after the interaction with groundwater,and its physical and mechanical properties also became poor,which was easy to produce extrusion deformation and led to compression and shear deformation of the sliding body towards the free direction,resulting in cracking of the rear edge and side boundary slope. Such cracks formed the main control structural plane. Before the landslide,the researchers installed two kinds of displacement monitoring sensors,the global navigation satellite system(GNSS) and the adaptive crack meter,to continuously monitor the deformation of the landslide. The on-site monitoring data is transmitted to the"geological disaster monitoring and early warning system" in real time,which was independently developed by researchers. The system automatically calculates the monitoring data and issues the early warning results through a variety of threshold comprehensive early warning models. According to the analysis of monitoring data,the landslide experienced the initial acceleration deformation stage,the intermediate acceleration deformation stage,the temporary sliding stage,and finally the disaster occurred. At each stage,the system released the early warning information in time according to the four-level early warning mechanism. Especially after the landslide entered the critical slide stage,the system sent out the red early warning 53 minutes in advance,which created the time condition for disaster emergency disposal and completely avoided the loss of personnel and economy. The successful early warning of the landslide reflected the applicability of the self-developed geological disaster monitoring and early warning system,early warning model and monitoring instrument,which can provide reference for research and application of similar landslide monitoring and early warning in the future.