Abstract: A time-varying fatigue reliability analysis method of steel bridges is proposed in the present study, which can consider the coupling effects of heavy loads and corrosion on the fatigue reliability of steel bridges. Taking a steel-concrete composite simply-supported bridge as an example, the investigated is the influence of average daily truck traffic, of equivalent gross vehicle weight and, of corrosivity grade on its fatigue failure probability. The calculated is the fatigue failure probability of its fatigue details under various working conditions, and the compared is the results of fatigue life calculated upon one deterministic analysis method and one reliability analysis method under various corrosion environments. The results show that: The proposed framework based on support vector regression and Monte Carlo simulation can calculate the fatigue failure probability of steel bridges quickly and accurately in corrosive environment; The increase of the average daily truck traffic, of equivalent total weight of vehicle and, of the level of environmental corrosivity will significantly shorten the fatigue life of steel bridges; The fatigue life estimated upon deterministic analysis tends to be conservative when the environmental corrosivity level is C5 and below. However, the results obtained upon the deterministic analysis may not be conservative when the environmental corrosivity reaches CX under which the environmental corrosion shall be fully considered.