Abstract: In order to study the mean square exponential stability of impulse stochastic delay neural networks with Markovian switching, a suitable Lyapunov function was established. Based on Itô formula, functional theory, the characteristics of neural networks and the knowledge of matrix, the state of system at impulsive time and non-impulsive time was analyzed, getting the rules to distinguish the mean square exponential stability of the stochastic neural network. The numerical example analysis and simulation results show that the rules are efficient, simple and validity.