Fire in road tunnels can lead to catastrophic consequences in combination with tunnel safety provision failures, thus necessitating a need for a reliable and robust approach to assess tunnel risks caused by fire. In a quantitative risk assessment (QRA) model for road tunnels, uncertainty is an unavoidable component because input parameters of the model possess different levels of uncertainties which are inappropriate to be formulated by crisp numbers. In this paper, a Monte Carlo sampling-based QRA model is proposed to address parameter uncertainty of a QRA model. The tunnel risks are assessed in terms of percentile-based societal risk as well as expected number of fatalities (ENF) curve, which would facilitate tunnel managers to make decisions. A case study is carried out to demonstrate the approach.