Previous　studies　have　confirmed　that　the　insulation　status　of　oil-paper　transformer　can　be　qualitatively　analyzed　by　its　time　domain　dielectric　spectral　characteristics.　However,　it　is　more　important　to　evaluate　the　insulation　condition　of　transformers　quantitatively　in　engineering　practice.　An　evaluation　system　based　on　characteristics　obtained　by　the　return　voltage　measurements　is　constructed　in　this　paper　for　the　quantitative　evaluation　of　the　insulation　state　of　oil-paper　transformers.　First,　data　of　characteristics　of　transformer　are　normalized　and　neighborhood　radius　of　each　characteristic　are　calculated.　Then,　the　decision　table　of　oil-paper　insulation　diagnosis　system　is　generated　based　on　neighborhood　rough　set　theory.　Moreover,　attribute　reduction　is　carried　out　to　the　decision　table　and　the　best　rules　are　extracted.　Finally,　the　oil-paper　insulation　state　evaluation　system　is　constructed　based　on　the　historical　database.　Five　transformers　of　known　insulation　status　have　verified　the　accuracy　of　the　evaluation　system.　©　2019　IEEE.