针对一类非线性组合大系统 ,提出一种用高阶神经网络逼近互联大系统的新型设计方法。首先用高阶神经网络逼近非线性组合大系统中的互联项 ,这样不仅可以解决大系统中最为复杂的互联项问题 ,且较以往采用的方法在工程上易于实现 ;然后基于高阶神经网络研究组合大系统的间接自适应控制问题。仿真结果表明了该方法的有效性。 For a class of nonlinear systems, a new design method for approximating interconnected large-scale systems with high-order neural networks is proposed. First of all, the high-order neural networks are used to approximate the interconnections in a nonlinear composite system. This not only can solve the most complex interconnected problems in large-scale systems, but also is easier to implement than previous methods. Then, based on the neural network of higher order Indirect Adaptive Control of Composite Systems. Simulation results show the effectiveness of the method.