In　order　to　solve　the　problem　of　measurement　between　vehicle　tires　and　highway　pavement,　this　paper　built　a　mathematical　model　among　axle　load,　tire　pressure　and　contact　interface　pressure　based　on　modular　neural　network　(MNN).　With　the　different　axle　load　and　tire　pressure,　the　membership　function　based　on　distance　measurement　is　constructed　firstly,　and　on　that　basis,　a　fuzzy　cluster　method　to　sample　data　is　adopted　to　realize　on-line　sub-nets　selection　to　improve　the　self-adapting　ability　of　MNN.　The　experiment　on　the　sample　data　of　Hefei　city＇s　highway　demonstrates　that　this　model　is　superior　to　others　in　accuracy　and　adaptability.