The invention discloses an oil-paper insulation system breakdown voltage prediction method based on a support vector machine. The oil-paper insulation system breakdown voltage prediction method comprises the steps of: (1) selecting a typical oil-paper insulation structure in a transformer, and forming an influence factor execution table; (2) conducting a breakdown voltage experiment according to the influence factor execution table, and recording breakdown voltage experimental data; (3) establishing an SVM regression prediction model; (4) separating a to-be-predicted oil-paper insulation system to establish an equivalent circuit model of each insulation structure; (5) calculating a breakdown voltage of each insulation structure by utilizing the SVM regression prediction model; (6) and comparing the breakdown voltage with a withstand voltage of the equivalent circuit model of each insulation structure to obtain a predicted breakdown voltage of the oil-paper insulation system. The oil-paper insulation system breakdown voltage prediction method utilizes the support vector machine to perform prediction analysis on the oil-paper insulation breakdown voltage, is simple and efficient, is conductive to solving the uncertainty in transformer oil-paper insulation analysis, reduces the amount of oil clearance breakdown tests in transformer oil and electrical feature study, and reduces test cost.