The　characteristics　of　graphene,　including　good　mechanical　properties,　low　density,　high　thermal　conductivity,　and　high　electron　mobility,　have　led　to　the　usage　of　graphene-filled　composites　in　various　fields.　So　far,　the　conductive　mechanism　of　graphene-filled　composites　is　controversial.　In　this　paper,　a　numerical　analysis　model　was　developed　based　on　RVE　(representative　volume　element)　theory,　where　the　macroscopic　model　was　equivalently　replaced　by　a　microscopic　model.　The　graphene　filler　was　modeled　as　a　two-dimensional　rectangle　that　was　positioned　and　angled　arbitrarily　without　intersecting　with　each　other　for　the　sake　of　numerical　calculation.　On　the　finite　element　model,　Ohm＇s　law　and　the　tunneling　effect　were　combined,　and　a　subroutine　was　constructed　for　real-time　estimation　of　the　distance　between　graphene　fillers　in　a　cell,　with　Ohm＇s　law　being　substituted　by　the　tunneling　effect　for　distances　less　than　3　nm.　The　electrical　conductivity　of　graphene-filled　composites　was　further　numerically　analyzed　based　on　the　influence　of　different　boundary　conditions.　The　numerical　analysis　results　were　in　good　agreement　with　the　experiments.　This　work　demonstrates　that　the　tunneling　effect　dominates　the　conduction　mechanism　of　conductive　particle-filled　composites　at　the　percolation　threshold　and　above　and　can　be　helpful　to　explain　the　response　mechanism　of　flexible　particle-filled　composite　sensors.　　©　2022　IEEE.