针对未知非线性系统提出了一种具有输出变化量加速度比例因子的无模型自学习控制方案，主要思想是利用BP 网络对系统输入输出量进行辨识，同时构造误差控制器，通过辨识学习器向误差控制器动态传递更新权阵，以实现对未知非线性系统的无模型误差自适应控制。Matlab7.0 平台下的仿真结果表明，控制效果良好。关键词： BP 网络；辨识学习器；误差控制器；自适应控制Abstract: One method of non-model error self-learning control with output variable acceleration proportion factor of non-linear system is presented. Its idea is that the input-output of the system is identified by BP networks and error controller is constructed simultaneity .The weight matrix of the error controller is dynamically adjusted by identifier to reality adaptive control. The system simulation example based on MATLAB7.0 shows this control method is effective.Keyword: BP network；identifier；error controller；adaptive control