The shear strength parameters (SSPs) of slip soil, which control the stability and evolution process of a landslide, exist some uncertainty that should be characterized and considered significantly in the stability analysis. In this paper, a typical reservoir landslide with double-sliding zones is taken as a case study. The statistical characteristics of SSPs in sliding zones are first analyzed using the geostatistical method and the Copula method, especially the spatial variation features, scales of fluctuation, and correlation of SSPs. Then, random variable models for the shallow sliding zone and nonstationary random field models with various vertical scales of fluctuation for the deep sliding zone are established separately. Finally, by employing the nonintrusive stochastic finite element program, the fluid–solid numerical simulation of a real hydrological year is carried out to obtain the landslide displacement, element failure probability, and factor of safety. The results demonstrate that incorporating spatial variability into the SSPs of slip soil led to a fundamental change in landslide stability and deformation. Our study provides some references for the reliability assessment of reservoir landslides considering the uncertainties and the spatial variation characteristics of SSP in the sliding zone.

控制滑坡稳定性和演化过程的滑动土的抗剪强度参数（SSPs）存在一些不确定性，在稳定性分析中应该对其进行表征和考虑。本文以典型的双滑区水库滑坡为例进行研究。首先利用地统计方法和Copula方法分析了滑动带SSPs的统计特征，特别是SSPs的空间变化特征、波动尺度和相关性。然后分别建立浅层滑动带的随机变量模型和深部滑动带不同垂直波动尺度的非平稳随机场模型。最后，通过使用非侵入式随机有限元程序，进行真实水文年的流固数值模拟，得到滑坡位移、单元失效概率和安全系数。结果表明，将空间变异性纳入滑移土的 SSP 导致滑坡稳定性和变形的根本变化。该研究为考虑滑带SSP的不确定性和空间变化特征的水库滑坡可靠性评估提供了一些参考。