Abstract: Rapid long run-out landslide is a sort of geo-hazard with catastrophic property, and the formational mechanism and the motive process of which are rather complex. Therefore, it has always been a sticky task in the field of landslid eresearching. In this regard, this paper takes an overall review on the achievements obtained in the last few decades upon the study of rapid long run-out landslide, including the mechanisms of it, the relative tests methods, and the numerical methods applying in simulation, respectively. The numerical simulation methods for rapid long run-out landslide totally fall into three categories based on landslide character and medium, including Discrete Medium Models, Continuum Medium Models, and Coupled Models, respectively. Among these models, typical Discrete Medium Models are based on Discrete Element Method(DEM),Discontinuous Deformation Analysis(DDA), and Numerical Manifold Method(NMM); while frequently used Continuum Medium Models are based on methods developed form the Hydrodynamics, such as Finite Difference Method(FDM),Finite Volume Method(FVM),Smoothed Particle Hydrodynamics(SPH),Arbitrary Lagrangian-Eulerian Method(ALEM),Cellular Automaton(CA),Lattice Boltzmann Method(LBM), and so on. In spite of the genres of numerical methods is various, each model has their own characteristics and applications. Generally speaking, Discrete Medium Models is suitable for the simulation of rockslides or rock avalanches, and Continuum Medium Models is suitable for the simulation of flow-like landslides. However, as rapid long run-out landslide is a phenomenon encompassing many complex motive characters, which cannot be reflected by simple block medium or fluid medium, therefore both Continuum Medium Models and Discrete Medium Models has their own limits in delineating the motive process of rapid long run-out landslide. To get rid of this dilemma, many scholars have developed variousCoupled Models with amore widely scope of application, which are combined by Continuum Medium Models and Discrete Medium Models, and it turn out to be aneffective approach and thus has gradually become a developmental tendency. In addition, the fact is widely acknowledged that the calculated quantity for rapid long run-out landslide in numerical simulation is astronomical, and it usually turns out to be time-consuming and difficult. It suggests that to develop and employ computing technology like GPU(Graphics Processing Unit) or other kind of parallel computation methods seems to be a good optionforthe solution of this problem. Presently, numerical methods like NNM,ALEM,CA, and FVM are still in the primary stage of their application in the simulation of rapid long run-out landslide, but these methods are very valuable in both theoretical and practical point of view. Therefore, this paper points out that duly effort should be distributed to the development and application of these methods. Still, many challenges remain unsolved in this field, like dynamic entrainment problems(the phenomenon that material on the path of moving takes away by the landslide) and air-blast caused by this kind of landslides, and the interaction between the sliding mass and the infrastructures. These aspects are still the weaknesses in the research of rapid long run-out landslide, and the tests and the numerical simulation methods in these aspects are still undeveloped, which need to be emphasized in the future.