Abstract: The radio communication signals usually turn out to be some fine character differences. In this paper, .after analyzing the instantaneous envelope of the frequency hopping signals which can be used as the fine characteristics of the individual communication transmitter, a method based on the instantaneous envelope of frequency hopping signals is presented to identify the individual frequency hopping radio according to the fine character differences. Firstly, the instantaneous envelope of the frequency hopping signals is extracted based on an improved wavelet transform algorithm of envelope extracting, the influence of noises is reduced. Secondly, the transient characteristics including the box dimensions and the information dimension are computed, the Instantaneous fine character of the hopping signals is transformed into a feature vector. Then, individual identification of the different frequency hopping radio based on the method of the Constructive Neural Network is realized. Finally, three different types of radio in real work are used to verify the validity, and the experimental results of the actual data have shown that our method is efficiency.