对于传统的神经网络中神经元模型在结构和信息存储能力上存在的不足 ,本文提出了一种基于广义小波基函数网络的神经元集聚模型。这种小波神经网络不仅收敛速度快 ,非线性逼近能力更好 ,而且具有内部结构变尺度、自适应调整和广义信息存储等智能化特点 ,更符合生物原型的实际情况。静态学习和准动态学习仿真实验证明这种神经网络结构的有效性。 In the traditional neural network, the neuron model has some shortages in the structure and information storage capacity. In this paper, a neural network clustering model based on the generalized wavelet basis function network is proposed. This wavelet neural network not only has the advantages of fast convergence speed, better nonlinear approximation ability, but also has the intelligent features such as internal structure scaling, adaptive adjustment and generalized information storage, and more in line with the actual situation of biological prototype. Static learning and quasi-dynamic learning simulation experiments prove the effectiveness of this neural network structure.