提出了一种新的模糊神经网络的快速参数学习算法 ,采用一些特殊的处理 ,可以用递推最小二乘法 (RLS)来调整所有的参数 .以前的学习算法在调整模糊隶属度函数的中心和宽度的时候 ,用的是梯度下降法 ,具有容易陷入局部最小值点、收敛速度慢等缺点 ,而本算法则可以克服这些缺点 ,最后通过仿真验证了算法的有效性 . A new fast parameter learning algorithm based on fuzzy neural network is proposed, which adopts some special processing to adjust all the parameters by Recursive Least Squares (RLS) .The previous learning algorithm can adjust the center of fuzzy membership function and Width, the gradient descent method is used, which is easy to fall into the local minimum, slow convergence and other shortcomings, while the algorithm can overcome these shortcomings, and finally verify the effectiveness of the algorithm through simulation.