Real traffic information was analyzed in the statistical characteristics and approximated as a Gaussian time series. A data source model, called two states constant bit rate (TSCBR), was proposed in dynamic traffic monitoring sensor networks. Analysis of autocorrelation of the models shows that the proposed TSCBR model matches with the statistical characteristics of real data source closely. To further verify the validity of the TSCBR data source model, the performance metrics of power consumption and network lifetime was studied in the evaluation of sensor media access control (SMAC) algorithm. The simulation results show that compared with traditional data source models, TSCBR model can significantly improve accuracy of the algorithm evaluation.