Abstract: In order to guarantee the quality of dynamic traffic data on the original data level, by simultaneous sampling with different sampling rates, the multi-detector dynamic system of urban expressway was constructed and the multi-detector dynamic system was analyzed by using wavelet transform. Then, the multi-scale traffic data fusion method based on wavelet transform and Kalman filter was proposed. Finally, validation and comparative analysis were carried out by using actual data measured from the north-south viaduct in Shanghai. Experiment results indicate that when the stochastic noise intensity is 2.5%, 5.0%, 7.5% and 10.0% respectively, the data fusion effects of the proposed method are superior to that of the compared methods.