In this paper, an $H_{\infty }$ estimation approach is given for an array of coupled stochastic complex networks with intermittent nonlinearity switching. A set of binary random variables are adopted to characterize the intermittent switching behavior of the involved nonlinearities. To effectively alleviate data collisions and save energy, the Round-Robin protocol is utilized to curb network congestions in data communication. For the coupled stochastic complex networks, we design a protocol-based $H_{\infty}$ estimator that not only resists stochastic disturbances, but also ensures the exponential mean square stability of the desired error system under a given disturbance attenuation level. With the help of the Lyapunov stability and convex optimization theories, sufficient conditions are provided for the expected estimator. Simulations are provided to illustrate the reasonability of our $H_{\infty }$ approach.