AbstractA robust assessment of tunnels due to uncertainties present in soil and ground motion properties can affect the dynamic response of these structures. In this paper, a stochastic analysis considering an aleatory variability in shear velocity Vs by performing Monte Carlo simulations and assessing its influence on underground tunnels. To numerically assess the response of the soil‐tunnel system to near‐fault earthquake ground motion, the required computational domain usually spans multiple kilometers including both earthquake fault and the structure. Thus, these simulations are computationally expensive. An efficient alternative method, that is, the Domain Reduction Method (DRM), in which a modular two‐step methodology for reducing the computational costs in the large domain analysis is employed. A 3‐D soil‐tunnel structure interaction is modeled to simulate the response of tunnel subjected to an inclined earthquake fault. The numerical simulation of the soil‐tunnel‐fault system is conducted in an open‐source FE package called Multi‐hazard Analysis for STOchastic time DOmaiN phenomena (MASTODON) based on MOOSE framework. The interactive analyses are carried out for three distinct types of soils based on National Earthquake Hazards Reduction Program (NEHRP) provision, that is, soft, medium, and hard, by generating 130 realizations. The results provide many insights into the influence of local site effects, that is, frequency shift for the peak response and the importance of input ground motion characteristics that govern the response of underground structures. The dynamic response of the structure is very sensitive to the uncertainty in rock material properties, especially, the stiffness of the rock sample and the shear wave velocity.