In　view　of　the　existing　harmonic　responsibility　division　methods　do　not　consider　the　actual　impact　difference　caused　by　different　harmonic　severity，a　long-time　scale　harmonic　responsibility　division　method　considering　the　harmonic　severity　is　proposed.　The　harmonic　value　distribution　and　trend　are　considered　to　divide　the　working　conditions，and　the　comprehensive　weight　of　each　working　condition　is　calculated　to　quantify　the　harmonic　severity.　The　long-time　scale　data　are　screened　based　on　the　principle　of　canonical　correlation　analysis，and　the　harmonic　responsibility　is　estimated　according　to　the　definition　of　harmonic　responsibility.　The　long-time　scale　comprehensive　harmonic　responsibility　division　index　is　obtained　by　combining　the　above　comprehensive　weights.　Simulation　examples　and　measured　data　are　used　to　verify　the　proposed　method.　Compared　with　the　traditional　method，the　proposed　method　can　reflect　the　cumulative　influence　of　different　harmonic　sources　in　a　long-time　scale，and　is　more　suitable　for　accurate　harmonic　control　and　fair　reward　and　punishment.　©　2024　Electric　Power　Automation　Equipment　Press.　All　rights　reserved.