As　an　essential　mode　of　last-mile　connectivity　for　public　transit,　dockless　bike-sharing　(DBS)　has　garnered　increasing　attention　in　the　analysis　of　feeder　trips.　However,　most　previous　studies　have　primarily　focused　on　land　use　attributes　around　stations,　neglecting　the　influence　of　factors　at　other　stages　such　as　trip　origins　and　route　environments.　To　address　these　gaps,　this　study　employs　XGBoost　and　SHAP　to　analyze　the　relationship　between　built　environment　attributes　and　DBS-metro　integrated　travel　at　origin-route-destination　level　based　on　multi-source　geographic　data　such　as　DBS　trajectory　data,　streetscape　images,　and　POIs.　The　results　indicate　that　route-built　environment　factors　have　a　stronger　influence　on　DBS-metro　integration　than　traditional　5D　attributes.　Furthermore,　the　influence　of　built　environment　factors　is　nonlinear.　When　the　green　view　index　is　between　0.15　and　0.25,　residents　are　attracted　to　using　DBS　to　reach　the　metro.　Moreover,　this　study　identifies　interaction　effects　between　cycling　distance　and　other　factors.　The　research　findings　provide　scientific　support　for　operators　to　allocate　vehicles　and　transportation　planners　to　undertake　community　regeneration　and　develop　sustainable　transportation　systems.