利用Lyapunov函数和线性矩阵不等式(LMI),给出了判定一类双向联想记忆(BAM)神经网络模型的指数稳定的充分性条件.该条件去掉了以往论文中所要求的激活函数单调,可微分的条件,而且所得结果利用里的工具易于检测.并举例说明本文结果的有效性. Using Lyapunov function and Linear Matrix Inequality (LMI), the sufficient condition of exponential stability of a class of bidirectional associative memory (BAM) neural networks is given, which eliminates the monotonous and differentiable functions of the activation function required in previous papers , And the results obtained are easy to detect using the tools in the article and illustrate the effectiveness of the results in this paper.