Urban　multimodal　public　transport　system　with　rail　transit　as　the　backbone　and　bus　as　the　collaborative　operation　is　an　important　carrier　for　the　travel　of　urban　residents.　For　improving　efficiency　of　urban　public　transport　services,　it　is　important　to　optimize　the　feeder　bus　operation　plans　and　promote　efficient　bridging.　A　quantitative　method　based　on　the　data　envelope　analysis　(DEA)　algorithm　was　proposed　to　effectively　evaluate　the　transfer　efficiency　of　transit　hubs　and　identify　service　bottlenecks,　which　used　multi-source　data　including　smart　card　transaction　data,　bus　vehicle　operation　data　and　public　transport　trip　chain　data　based　on　multimodal　individual　travel　transaction　data.　Considering　the　passenger　transfer　and　supply　capacity　of　the　feeder　bus,　and　balancing　between　transfer　and　regular　travel　needs,　a　collaborative　optimization　model　based　on　the　branch　and　bound　(BB)　algorithm　which　can　quickly　obtain　the　absolute　optimal　solution　of　small-scale　optimization　problems　is　established　to　combinatorial　optimize　less　efficient　feeder　bus　line　groups.　The　Tiantongyuan　North　transit　hub　and　its　feeder　bus　line　in　Beijing　were　selected　as　the　case　analysis.　The　results　indicated　that　the　main　reasons　for　the　failure　of　public　transport　to　achieve　DEA　effectiveness　are　the　departure　frequency　and　the　number　of　operating　buses.　With　the　collaborative　optimization　of　the　bus　routing　schedule,　the　average　total　transfer　time　for　the　transfer　passengers　is　reduced　by　19.9%,　while　the　transfer　time　and　waiting　time　are　reduced　by　22.8%,　16.1%　respectively,　which　will　help　improve　the　efficiency　and　service　level　of　multimodal　public　transport　collaborative　operation.　©　2024　The　Author(s)