A　simulation　model　is　established　based　on　Jackson＇s　algorithm,　which　captures　five　social　networks＇　properties,　to　analyze　the　dynamic　characteristics　of　tacit　knowledge　diffusion　in　this　paper.　Different　local-level　tacit　knowledge　exchange　behavioral　patterns　are　considered　in　this　model,　namely　＇tacit　knowledge　trade＇　and　＇tacit　knowledge　teaching＇.　The　steady-state　knowledge　level　and　the　process　of　tacit　knowledge　diffusion　evolution　under　different　social　network　topology　and　tacit　knowledge　exchange　patterns　are　simulated　and　analyzed.　The　simulation　results　demonstrate　that　the　efficiency　of　tacit　knowledge　diffusion　decrease　as　the　social　network　heterogeneity　increase.　And　the　＇tacit　knowledge　teaching＇　pattern　can　significantly　improve　the　efficiency　of　tacit　knowledge　diffusion.