Abstract: Traffic measurement and modeling is essential to planning and management of existing IP networks，as well as to designing next generation networks.In this paper，a new P2P traffic prediction model based on wavelet transform and times series analysis is introduced.The original traffic metric series are decomposed and reconstructed into several branches by wavelet.These branches，stabilized by a statistic method，are predicted by ARIMA or ARMA models respectively according to their different features.The final predicted results can be obtained by reconstructing these predicted values of branches and an optimizing method with the use of dynamic exponential smoothing model.Experimental results show that forecasting trends are very similar with the actual flow curves，and the Mean Relative Error（MRE） is around 4.04%，more accurate than the traditional methods.