To　understand　the　travel　mode　choice　characteristics　of　passengers　arriving　at　transportation　hubs　and　address　the　optimization　and　efficiency　improvement　of　transport　capacity　allocation,　data　from　high-speed　rail　hub　travel　questionnaires　are　analyzed.　Latent　variables　including　the　attributes　of　previous　connections,　the　subjective　importance　of　connection　modes,　hub　service　satisfaction,　and　travel　choice　tendencies　are　introduced.　A　Structural　Equation　Model　(SEM)　that　considers　psychologi⁃　cal　tendencies　is　constructed　and　integrated　with　a　Multinomial　Logit　(MNL)　model,　creating　a　com⁃　bined　SEM-MNL　model.　This　model　quantitatively　calculates　the　utility　functions　for　different　travel　mode　choices　of　arriving　passengers,　predicts　the　probability　of　choosing　various　modes　of　travel　at　dif⁃　ferent　times,　and　compares　the　hit　rate　of　the　SEM-MNL　model　with　that　of　the　MNL　model.　The　re⁃　sults　show　that　incorporating　psychological　tendency　variables　improves　the　overall　hit　rate　of　the　model　by　4.6%.　Passengers　with　higher　satisfaction　with　hub　services,　greater　importance　placed　on　connection　services,　shorter　travel　time　and　distance　in　their　previous　trip,　and　those　using　public　transport　are　more　likely　to　choose　public　transport　modes.　Predicting　travel　mode　demand　using　the　mode　choice　model　and　ensuring　transport　capacity　coordination　at　passenger　hubs　is　crucial　for　balanc⁃　ing　supply　and　demand　between　passenger　flow　and　capacity　resources,　thereby　improving　service　quality.　©　2024　Journal　Northern　Jiaotong　University.　All　rights　reserved.