As a natural geologic disaster, soil landslide is affected by natural factors and human activities. It brings a great threat to the safety of the surrounding residents' life and property, attracting more worries from people. Landslide prevention and control are also one of the hot spots in engineering research. Soil is essentially a kind of deposits consists of granular material with complicated composition and structure. By means of numerical simulation of granular material flows, we can well understand the phenomena of soil flows in the nature, such as landslide and debris flows. It is therefore possible to predict the landslide hazard zone and to improve the design of engineering protection measurements. However, the finite element method(FEM) is not suitable for simulating soil flow with large deformation and large displacement due to the finite element method is sensitive to mesh distortion. In this paper, we simulate the soil flows using the material point method(MPM). MPM is a mesh-free method and originating from the particle-in-cell(PIC)method. It combines the Eulerian description and Lagrangian description and has distinct advantages in solving the large deformation and large displacement problem. Currently, there have been many studies on the simulation of slope sliding using MPM,but less attentions has paid to sensitivity analysis on relevant parameters. In this paper, the large deformations of the cohesive soil and the non-cohesive soil slopes due to gravity are numerically investigated with MPM.The influence of the internal friction angle, the cohesive force, the aspect ratio and bottom surface gradient on the landslide run-out are analysed. In this MPM simulation, the elasto-palstic constitutive models based on the Drucker-Prager(DP)yield criterion is used for modeling nonlinear characters of soil flows. The Drucker-Prager yield criterion is a pressure-dependent model for determining whether a material has failed or undergone plastic yielding. The yielding surface of the Drucker-Prager criterion may be considered depending on the internal friction angle of the material and its cohesion. The simulated results show the following featurs. (a) There are well agreements in the flow pattern, the sliding distance and natural angle of repose between the simulated results and the corresponding experimental results, which validates the ability of MPM to modeling soil flows. (b) The sliding distance decreases with the increase of the internal friction angle and cohesive force. (c) The steepness of the soil slope has significant influence on its stability. The sliding distance increases with the increase of the steepness of the soil slope. (d) For a comparatively small aspect ratio, the sliding distance increases linearly with aspect ratio. It is noteworthy that the internal friction angle and cohesive force of soil reflect its shear resistance. Therefore, the damage due to landslide can be reduced by improving the shear resistance of soil through engineering measurements. The results of the simulation provide a reliable reference to explore the law of the soil sliding hazardous behavior and reduce the sliding damage range.