The present application provides a method for identifying pre-existing structures controlling a porphyry metallogenic system based on a neural network model, falling within the technical field of magmatic hydrothermal ore deposit exploration. The method includes:obtaining structural evolution sequence of all mining areas, the sequence of tectonic-metallogenic rock mass-ore body formation, and the key identification characteristics signs of pre-existing structures. The present application substitutes geological data of the mining area of a porphyry metallogenic system to be identified into the trained pre-existing structure identification model to identify the pre-existing structure, simplifying identification process, improving accuracy, reducing subjective influence caused by artificial evaluation, and improving characterization ability of the identification result.