The　polishing　efficiency　and　accuracy　for　optical　elements　are　strongly　affected　by　the　removal　characteristics　of　the　polishing　tool　and　are　commonly　evaluated　using　a　tool　influencing　function(TIF).　During　most　polishing　processes,　the　TIF　is　obtained　through　offline　measurements,　which　are　time-consuming.　Consequently,　the　use　of　polishing　forces　collected　online　to　evaluate　TIF,　including　its　spot　shape　and　efficiency,　was　investigated　as　a　means　of　reducing　measurement　times.　Based　on　the　elastic　theory,　the　formation　mechanism　for　an　elliptical　TIF　spot　during　a　bonnet　polishing　process　is　thought　to　have　a　long　axis　length　that　is　close　to　a　predefined　value　and　a　short　axis　length　that　is　proportional　to　the　third　root　of　the　normal　force.　Based　on　the　above　finding,　a　polishing　force-based　prediction　model　for　TIF　spot　shape　was　derived　and　validated.　The　predicted　results　agreed　well　with　experimental　results.　Afterwards,　a　polishing　force-based,　semi-quantitative　prediction　model　for　TIF　efficiency　was　developed　by　combining　Preston＇s　law　and　Hertz　contact　theory.　The　model　was　validated　experimentally　and　successfully　predicted　changing　trends　in　TIF　efficiency.　Based　on　this　study,　the　spot　shape　and　efficiency　of　the　TIF　can　be　predicted　using　polishing　forces　collected　online.　This　approach　provides　a　potential　method　that　could　be　used　in　engineering　applications　to　optimize　bonnet　polishing　processes　through　a　reduction　in　the　time　needed　for　offline　measurements.