This　paper　proposes　a　land　use　inferring　method　based　on　the　convolutional　neural　network　(CNN),　which　can　infer　multiple　lane　use　types　at　the　traffic　analysis　zones　(TAZs)　simultaneously.　The　study　combines　public　transport　mobility　dataset　and　online　car-hailing　mobility　dataset　for　inferring　land　use　type.　Generation　intensity,　attraction　intensity,　and　difference　between　generation　and　attraction　intensity　are　extracted　from　the　travel　dataset,　which　are　then　used　to　train　the　CNN.　The　optimal　network　structure　is　determined　by　grid　search.　The　TAZs　within　the　6th　Ring　Road　of　Beijing　are　taken　as　examples　for　the　analysis.　The　results　indicate　that　the　proposed　method　is　able　to　estimate　the　proportion　distribution　of　several　land　use　types　at　the　same　time　within　the　TAZs,　such　as　resident,　workplace　and　leisure　land　uses.　Copyright　©　2020　by　Science　Press.