Constructing　an　effective　multi-level　public　transport　system　is　an　efficient　way　to　meet　the　differential　travel　demands　of　travelers　and　support　sustainable　development　in　urban　areas.　Based　on　the　analyses　of　trip　characteristics　of　various　modes　of　public　transport,　an　ordered　logit　model　of　willingness　to　pay　(WTP)　was　constructed　and　the　service　level　factors　of　travelers＇　WTP　were　analyzed.　A　K-means　clustering　method　as　used　to　isolate　three　trip　groups,　the　WTP　elasticity　of　various　traveler　types　was　analyzed　and　the　differences　in　WTP　between　these　trip　groups　were　studied.　Finally,　a　virtual　scene　was　constructed　using　differentiated　bus　service　and　the　WTP　of　travelers　for　the　scenario　was　calculated.　The　results　of　the　scenario　showed　that　travelers＇　WTP　for　bus　service　is　largely　affected　by　speed,　reliability　and　degree　of　crowding.　The　most　important　measures　by　which　to　improve　the　attractiveness　of　bus　service　were　identified　to　be　both　speed　and　reliability.　©　2018　American　Society　of　Civil　Engineers　(ASCE).　All　rights　reserved.