The　author　attempts　to　do　research　on　effect　factors　related　to　proportion　of　transit　trips　first　in　this　paper.　BP　neural　network　prediction　model　is　established,　based　on　available　city　traffic　survey　data.　Meanwhile,　we　develop　corresponding　programs　by　taking　advantage　of　neural　network　toolbox　in　MATLAB.　It　is　evident　that　it　has　characteristics　of　high　learning　capability,　accurate　prediction　and　convenience.　And　it　provides　a　thinking　and　method　of　proportion　of　transit　trips　prediction　research　as　a　reference.