The　uncertainty　of　travel　information　was　described　by　using　interval　numbers,　and　a　travel　mode　choice　model　with　multiple　attributes　was　put　forward.　The　least-square　method　based　on　complementary　judgment　matrix　was　used　to　obtain　the　weights　of　different　travel　modes,　the　weight　of　each　attribute　variable　was　computed　based　on　a　dispersion　maximization　model,　and　the　composite　weight　of　each　travel　mode　was　obtained.　The　efficiency　of　the　choice　model　was　validated　according　to　car　and　public　transit　travel　samples.　Analysis　result　shows　that　car　is　usually　chosen　when　its　travel　weight　exceeds　0.　65,　and　public　transit　is　usually　chosen　when　its　travel　weight　exceeds　0.　68,　so　the　model　can　simulate　the　travel　mode　choices　of　travelers　effectively.　1　tab,　2　figs,　9　refs.