<p>Determining post-failure characteristics of slopes under rainfall is an essential prerequisite for assessing and mitigating landslide risks. This study proposes a coupled hydro-mechanical framework that integrates the Random Limit Equilibrium and Material Point Methods (RLE-MPM). The post-failure characteristics, landslide types and probability of slope failure considering the spatial variability of saturated hydraulic conductivity (<i>k</i>_<i>s</i>) are evaluated using this framework through probabilistic large deformation analyses. An infinite slope under rainfall and Tokai-Hokuriku expressway landslide in Japan are investigated, respectively, to illustrate the effectiveness of the proposed RLE-MPM framework. The influences of the rainfall duration as well as the coefficient of variation of mechanical parameters and <i>k</i>_<i>s</i> on the slope post-failure characteristics are investigated. Results indicate that the proposed RLE-MPM framework is computationally more efficient in evaluating the slope post-failure characteristics and probability of slope failure than the Random Material Point Method (RMPM). Additionally, the spatial variability of <i>k</i>_<i>s</i> significantly affects the means and coefficients of variation of slope post-failure characteristics. The failure characteristics of the Tokai-Hokuriku expressway landslide can be accurately predicted by using the proposed framework. This study highlights the necessity of considering the heterogeneity of soils in landslide movement prediction and risk assessment.</p>