The　insulation　state　evaluation　of　transformer　oil　paper　is　a　complex　problem　with　multiple　features.　In　this　paper,　an　improved　analytical　hierarchy　process　and　technique　for　order　preference　by　similarity　to　ideal　solution　(AHP-TOPSIS)　method　of　insulation　state　evaluation　model　with　positive　and　negative　reference　points　is　proposed.　First,　the　response　voltage　method　(RVM)　is　utilized　to　extract　the　related　features　of　the　insulation　state　of　oil　paper.　The　hierarchical　structure　of　the　insulation　state　evaluation　of　the　transformer　oil　paper　is　formulated.　Secondly,　the　analytic　hierarchy　process　is　applied　to　determine　the　weight　coefficients　based　on　the　characteristic　quantity　of　the　expert　experience.　Then,　the　state　is　divided　into　the　fine　insulation　and　the　serious　aging　degradation.　These　two　factors　are　adopted　as　references　to　assess　the　transformer.　In　this　way,　a　decision　matrix　can　be　constructed,　which　is　normalized　and　combined　with　the　weight.　The　positive　and　negative　reference　points　of　this　paper　are　positive　and　negative　ideal　solutions.　Finally,　the　close　degree　of　the　transformer　and　the　positive　reference　point　are　calculated.　The　insulation　state　of　transformer　is　evaluated　according　to　the　interval　of　proximity.　The　application　example　proves　the　accuracy　of　the　improved　AHP-TOPSIS　method　in　the　insulation　state　evaluation　of　the　transformer.　Compared　with　the　traditional　TOPSIS　method,　the　insulation　state　of　the　transformer　oil　paper　can　be　evaluated　more　accurately.　©　2018,　Science　Press.　All　right　reserved.