Aiming　at　traditional　method　of　parameter　identification　for　relaxation　response　equivalent　circuit,　which　can　not　eliminate　fluctuations　of　charging　voltage　and　measurement　error,　combined　with　depolarizing　current　data,　the　interval-adaptive　particle　swarm　optimization　(I-APSO)　was　introduced　to　identify　the　parameters　of　oil-paper　insulation.　The　algorithm　not　only　reduces　the　actual　error　making　the　result　more　precise,　but　also　locates　the　optimal　solution　areas　of　the　parameters　effectively.　The　radical　exponential　decay　law　was　used　to　improved　dielectric　response　function　considering　that　the　existing　one　can　not　reflect　the　medium　depolarization　process　accurately.　The　result　shows　that　depolarizing　current　curve　by　I-APSO　and　improved　dielectric　response　model　can　be　matched　better　with　test　curve　compared　with　conventional　method,　which　reflects　the　dielectric　process　of　oil-paper　insulation　system　more　truly　and　effectively.　©,　2015,　Editorial　Department　of　Electric　Machines　and　Control.　All　right　reserved.