Abstract: The failure plane is a critical part of the entire landslide components and plays a key role in the form, development and evolution of a landslide. Therefore, the determination of the shear strength parameters at the failure plane is very important in the landslide stability evaluation. However, it is difficult to accurately obtain the design parameters due to the discrete nature of the failure plane distribution, the sampling disturbance and the potential testing errors. For landslide with complicated sliding bodies, the strength parameters obtained from the back analysis with the three-dimensional numerical analysis is still a good method, although the many disadvantages exist in the trial-and-error method in the numerical analysis. This paper presents a case history of three-dimensional numerical back analysis results on a landslide at the Yunnan Tengchong Airport using Particle Swarm Optimization and Differential Evolution approach. The detailed numerical model was established based on the actual topographical conditions and the strength parameters at the failure plane were obtained. This method significantly improves the efficiency of the back analysis process.