The relation model established by artificial neural network that includes 4 corrosive material parameters that are Cl -, H 2 S, NH 3 and pH value and corrosion testing parameter that is Fe 2+ of the fractionator overhead recycle system, studied the corrosion sensitivity of the corrosion parameters, has got the sensitive areas, put forward the suitable range of the corrosive material parameters of corrosion control in the production process. The corrosion of catalytic fractionator overhead recycle system has become one of the main factors in the refinery operating smoothly and stable production, not only cause the recycle system equipment damage, shorten the cycle of system safety operation, but also reduce the product quality of the subsequent processing device [1]. The corrosion factors of the fractionator overhead recycle system and their relationships are very complicated, and the combination of corrosive material parameters is diversification, it is very essential that research the corrosion model of the fractionator overhead recycle system and the corrosive material parameters' sensitivity.