The　structure　of　a　power　transformer　through　the　combination　of　oil　and　paper　as　the　main　insulation　system　is　the　key　to　its　safe　and　reliable　operation.　In　this　paper,　a　grey　correlation　model　is　constructed　by　Support　Vector　Machine　(SVM)　from　the　perspective　of　the　electric　field　change　during　the　discharge　process,　and　the　model　is　optimized　by　using　the　idea　of　cross-validation　to　find　the　optimal　parameters　to　construct　the　optimal　transformer　oil　gap　frequency　breakdown　voltage　prediction　model.　In　this　paper,　the　SVM　model　is　used　to　predict　the　transformer　oil　gap　frequency　breakdown　voltage　from　the　30　min　withstand　voltage　test,　and　the　predicted　breakdown　voltage　is　compared　with　the　actual　breakdown　voltage.　The　validity　of　the　model　was　verified.　©　2023　IEEE.