The　sample　adaptive　offset　(SAO)　filter　introduced　in　the　high　efficiency　video　coding　standard　can　remove　the　ringing　artifacts　caused　by　the　loss　of　high-frequency　information.　But,　it　dominates　the　complexity　of　in-loop　filtering.　In　this　article,　an　algorithm　optimization　and　implementation　scheme　with　low　complexity　SAO　is　proposed.　The　information　statistical　range　is　optimized　regionally,　which　avoids　the　complex　logic　control　across　coding　tree　unit　boundary,　and　saves　additional　cache　structure.　To　trade　off　the　complexity　of　SAO　mode　decision　and　the　accuracy　of　rate　model　in　rate-distortion　optimization,　a　three-dimensional　linear　rate　estimation　model　for　fast　calculation　of　rate-distortion　cost　is　proposed.　Finally,　based　on　the　above　improved　algorithms,　a　SAO　hardware　filter　suitable　for　ultra　high　definition　(UHD)　video　processing　is　designed.　The　proposed　architecture　supports　the　UHD　video　processing　up　to　333　MHz　with　only　85.　6　k　equivalent　gates.　Compared　with　the　traditional　scheme,　the　logical　resource　consumption　can　be　saved　up　to　71.　47%　.　©　2023　Science　Press.　All　rights　reserved.