Aimed　to　improve　the　modeling　accuracy　of　tool　influence　function　(TIF)　of　bonnet　polishing,　a　theoretical　and　experimental　study　is　presented.　This　paper　starts　with　the　affecting　mechanism　of　key　parameters　on　the　material　removal　of　workpiece.　It　is　indicated　that　the　interfacial　friction　coefficient　between　tool　and　workpiece　is　changed　with　the　variety　of　the　tool　rotational　speed,　which　impacts　the　TIF　but　has　not　been　taken　into　account　in　most　current　TIF　models.　Consequently,　modification　of　TIF　model　based　on　the　interfacial　friction　coefficient　is　proposed　and　then　experimentally　validated.　The　results　show　that,　for　the　experimental　groups　in　which　the　spot　size　is　15　mm,　the　difference　between　the　maximum　removal　depth　of　the　TIF　predicted　with　the　pre-modified　model　and　that　of　the　experimental　TIF　is　-0.204　to　1.244.　(lambda　=　632.8　nm),　which　is　obviously　larger　than　that　between　the　TIF　predicted　with　the　modified　model　and　the　experimental　TIF　-0.187　to　0.168..　Moreover,　for　the　experimental　groups　in　which　the　spot　size　is　20　mm,　the　difference　between　the　maximum　removal　depth　of　the　TIF　predicted　with　the　pre-modified　model　and　that　of　the　experimental　TIF　is　-0.135　to　2.235.,　while　that　between　the　TIF　predicted　with　the　modified　model　and　the　experimental　TIF　is　-0.046　to　0.571..　The　experimental　results　indicated　that　the　TIF　predicted　by　the　modified　model　is　much　closer　to　the　experimental　TIF,　which　proves　the　effectiveness　and　correctness　of　the　modification.