Most of the existing infrastructures (e.g., oversea bridges) are made of reinforced concrete (RC); therefore, their safety is of high priority for our society. RC structures undergo continuous deterioration due to a combination of chloride ingress and loading actions. Chloride-induced corrosion is thus one of the most serious threats. In this work a stochastic model is presented which couples degradation effects of crack development and corrosion progression based on physicochemical and mechanical models to estimate the probability of failure of a RC structure over time. To enable an efficient probabilistic assessment, a novel rapid numerical approach (RNA) is implemented for modeling the chloride diffusion. Numerical experiments show that the probability of corrosion failure is significantly underestimated if the coupled degradation effect of cracks and corrosion is not considered. Finally, parameter analyses are used to identify the effects of individual parameters, including external load level, supplementary cementitious materials and environmental temperature, on the probability of corrosion failure of RC beams over time.