In this paper, the performance-based control design problem for double-layer networked industrial processes is investigated. At the device layer, the prescribed performance functions are first given to describe the output tracking performance, and then by using backstepping technique, new adaptive fuzzy controllers are designed to guarantee the tracking performance under the effects of input dead-zone and the constraint of prescribed tracking performance functions. At operation layer, by considering the stochastic disturbance, actual index value, target index value, and index prediction simultaneously, an adaptive inverse optimal controller in discrete-time form is designed to optimize the overall performance and stabilize the overall nonlinear system. Finally, a simulation example of continuous stirred tank reactor system is presented to show the effectiveness of the proposed control method.