The identification of chaos phenomenon is important for chaos analysis and control.Based on the conceptions and methods of information theory,control theory and data mining techniques,the framework and function module of the system for real-time identification of chaos are introduced.To meet the requirement of the principle and function of this system,feature extraction for time series is necessary.The case studies are given for time series generated by Logistic system.Feature of time series are extracted by wavelet packet coefficients and wavelet packet energy respectively.The reason and necessity of feature extraction are analyzed.The results indicate that the module for feature extraction is a key module in the system for real-time identification of chaos.Wavelet packet energy can be used in this system for feature extraction.