在小波神经网络的基础上提出了一种辨识非线性动态系统的方法．该方法有效地将系统辨识所需要的结构形式与多层神经网络及小波基函数所构成的分辨率信息处理过程相结合，建立了从数据到符号的自适应机制．仿真结果表明，该方法具有收敛速度快、逼近精度高、鲁棒性好等优点． Based on the wavelet neural network, a method of identifying nonlinear dynamic system is proposed. The proposed method effectively combines the structure required by system identification with the resolution information processing of multi-layer neural networks and wavelet basis functions, and establishes an adaptive mechanism from data to symbols. Simulation results show that the proposed method has the advantages of fast convergence, high approximation accuracy and good robustness.