In　order　to　solve　the　problem　that　the　optimal　decision　obtained　by　the　traditional　reactive　power　optimization　model　is　difficult　to　adapt　to　the　voltage　over-limit　caused　by　the　ultra-short-term　forecast　deviation　of　wind　power　in　the　actual　scheduling,　a　reactive　power　adjustment　strategy　for　loss　reduction　considering　the　ultra-short-term　forecast　deviation　of　offshore　wind　power　is　proposed.　The　probability　density　function　of　voltage　deviation　is　obtained　based　on　Gaussian　mixture　model　and　linearized　power　flow　equa-tion　in　order　to　evaluate　the　nodes　with　the　risk　of　voltage　over-limit　and　the　voltage　fluctuation　ränge　under　the　forecast　deviation　of　active　power　Output.　Based　on　this,　combined　with　the　influence　mechanism　of　the　forecast　deviation　of　active　power　Output　on　voltage,　an　optimization　model　of　reactive　power　adjustment　for　loss　reduction　is　established　to　reduce　the　reactive　power　optimization　ränge　of　the　corresponding　wind　turbines　based　on　the　maximum　voltage　deviation　of　the　risk　node　under　the　influence　factor　of　a　certain　risk　node,　so　as　to　improve　the　safety　and　economy　of　the　reactive　power　adjustment　strategy　for　loss　reduction.　A　real　91-bus　offshore　wind　farm　is　taken　as　a　Simulation　example　to　verify　that　the　proposed　method　can　avoid　high　voltage　over-limit　penalty　costs　to　the　maximum　extent,　and　obtain　high　loss　reduction　benefits　by　reactive　power　adjustment.　©　2025　Electric　Power　Automation　Equipment　Press.　All　rights　reserved.