This paper presents a design of a self-organizing recurrent interval type-2 Petri fuzzy controller (SORIT2PFC) for controlling the time-varying delay systems. The control system comprises of a main controller and a compensation controller; the SORIT2PFC as the main controller is used to imitate an ideal controller, and the simple fuzzy compensator controller is used to eliminate the residual error. By using a self-organizing algorithm, the structure of the proposed network can automatically achieve optimal construction. Recurrent neural network and fuzzy Petri nets are applied to improve system performances and to reduce computational burden. Online tune adaptive laws of the proposed controller are derived by implementing the gradient descent method and the Lyapunov stability theorem. Finally, the control efficacy and effectiveness of the proposed controller are verified by the numerical simulations of the time-varying delay systems.INDEX TERMS Type-2 fuzzy system, Petri-fuzzy, recurrent neural network, self-organizing algorithm, time-varying delay systems.