Highway bridges are structures subjected to the both action of fatigue and corrosion. In this sense, this research proposes the prototype of a system to assist in the maintenance decision making process of steel beam bridges, in which only simply supported steel girder bridges are considered. The system contemplates the effects of corrosion-fatigue and estimates the lifetime of these structures based on a structural reliability analysis. Such analysis is based on AASHTO S-N fatigue curves and Miner s damage accumulation rule. In order to account for material and resistance loss caused by corrosion, adjustments are made to the fatigue parameters of the reliability model used. In addition, some maintenance strategies that reduce the rate of corrosion are considered in the system by correcting the corrosion parameters of the material. The system uses an artificial neural network model to estimate the stresses in steel beams under the passage of an AASHTO fatigue vehicle using geometric and material data from the bridge. The database used for the development of the neural network was created from finite element simulations results. The life cycle of a steel bridge, designed according to the AASHTO American standard, is simulated using the proposed system. Through this simulation, it is concluded that the structure s lifetime is related to the combination of traffic parameters and the environment corrosivity. It is also observed that the increase in the average daily truck traffic (ADTT) can cause a reduction from 48 percent to 76 percent in the bridge s lifetime and depending on its annual growth increase rate the reduction can go up to 80 percent. Maintenance alternatives for the superstructure such as washing provide a lifetime gain for all environments, especially in the marine environment (up to 30 percent). However, in some of the tested traffic scenarios this gain is not sufficient to guarantee the minimum service life recommended by AASHTO.