本文对具有不确定性控制对象提出了一种自学习模糊神经网络控制方法模糊控制器采用误差，误差变化及误差加速度的加权和的解析描述形式，利用人工神经网络直接对过程的建模，实现对模糊加权因子的自学习优化调整将上述方法用于焊接熔池动态过程控制实验，结果表明本文提出的自学习模糊神经网络控制方案有效 In this paper, we present a self-learning fuzzy neural network control method for uncertain control objects.  The fuzzy controller uses the weighted sum of errors, error changes and error accelerations to describe and describe the model. The artificial neural network is used to directly process, Realization of the self-learning optimization adjustment of fuzzy weighting factor  The above method is applied to the experiment of weld pool dynamic process control. The results show that the proposed self-learning fuzzy neural network control scheme is effective