Summary: This paper is devoted to investigating global robust exponential stability for a class of delayed stochastic reaction-diffusion recurrent neural networks. The network parameters are governed by a continuous-time discrete-state Markov process which takes values in a finite set. By employing a Lyapunov-Krasovskii functional and some inequalities, some easy-to-test criteria on global exponential stability for this kind of stochastic neural networks are established in the form of linear matrix inequalities. An example is presented to illustrate the effectiveness of the theoretical results.