Abstract: In practical engineering, due to the slope stratification or the spatial variability of soil properties, the slope will often fail along multiple potential sliding surfaces. Considering that using a single critical deterministic sliding surface or a critical probability sliding surface to analyze slope reliability will greatly underestimate the probability of failure, this study uses the weighted uniform simulation (WUS) method to analyze the system reliability of slopes. Four multi-layer slope cases are used to prove that the WUS has good applicability for system reliability analysis of slopes. High accuracy is obtained for multi-layer slope problems under high-dimensional and implicit limit state equations by this method, while the sample size is greatly reduced from 10⁴ samples required by the direct Monte Carlo simulation to only about 500 samples under the same accuracy requirements. By modifying the original WUS algorithm, when the sample size is sufficient, the modified WUS can efficiently obtain the system failure probability and effectively abandon redundant failure modes to obtain accurate most probable failure points corresponding to multiple failure modes. Furthermore, the representative sliding surfaces of the multi-layer slope can be effectively and automatically identified, which provides important reference value for the later maintenance and failure prevention of slopes.