Probabilistic slope stability assessment offers a more realistic treatment of uncertainties in geotechnical parameters and modelling assumptions. This study presents a practical procedure for probabilistic analysis with relatively low computational cost by combining the Response Surface Method (RSM) with the First-Order Reliability Method (FORM). The approach is applied to a slope composed of highly weathered weak rock mass located on the Northern Ring Road in São Paulo, Brazil. Numerical simulations were performed using the Shear Strength Reduction Method within the Finite Element Method, employing average parameter values, while cohesion and friction angle of the dominant discontinuity were systematically varied according to a Central Composite Rotatable Design. Response surfaces derived from this design were integrated with FORM to estimate the probability of failure and identify the most probable combination of shear strength parameters associated with slope instability. Results from the probabilistic analysis were compared with deterministic back-analyses based on conventional factors of safety, highlighting the limitations of relying solely on deterministic approaches and illustrating the practical utility of FORM in capturing parameter uncertainties. The comparison demonstrates that FORM provides accurate and computationally efficient probabilistic predictions, serving as a robust tool for engineering decision-making in geotechnical practice.
Avenida República do Líbano, 251 - Pina, Recife - PE - CEP: 51110-160