研究了用神经网络建立非线性动态系统的数学模型。由于实际系统固有的非线性、复杂性及不确定性 ,建立系统的模型通常是困难的。文中讨论了四类非线性离散系统的建模方法 ,并将其应用到了润滑油的溶剂脱蜡生产过程 ,得到了令人满意的结果 The mathematical model of establishing nonlinear dynamic system by neural network is studied. Due to the inherent non-linearity, complexity and uncertainty of the actual system, it is often difficult to build a systematic model. In this paper, the modeling methods of four kinds of nonlinear discrete systems are discussed and applied to the solvent dewaxing process of lubricating oil, and satisfactory results are obtained