An　urban　impedance　computing　model　was　proposed　based　on　the　points　of　interest　from　social　networks　for　real-time　calculation　of　urban　impedance.　The　model　used　the　land　blocks　divided　by　urban　traffic　networks　as　the　basic　units,　and　factors　such　as　the　average　traffic　distance,　turning　angle,　turning　times,　intersection　and　the　population　and　so　on　were　considered　to　calculate　the　urban　impedance　within　an　accessible　radius.　In　the　experiment,　the　urban　impedance　was　calculated　for　each　2　h　using　the　proposed　model,　and　was　compared　with　the　utility　model　and　the　gravity　model.　The　results　show　that　the　similarity　between　our　impedance　model　and　the　actual　road　situation　is　80%　higher　than　the　utility　model　and　60%　higher　than　the　gravity　model,　respectively.　The　proposed　model　improves　the　applicability　of　urban　impedance　calculation　and　reflects　the　variation　in　different　time　periods.　©　2016,　Zhejiang　University　Press.　All　right　reserved.