This paper presents a neuro sliding mode controller for the twin rotor control system (TRCS). The TRCS belongs to a class of nonlinear systems having unstable coupled dynamics. A method is proposed to design a switching surface in the presence of mismatched uncertainties. The nonlinearities of the TRCS are assumed to be unknown, which are estimated using Chebyshev neural network (CNN). The design technique is simple and computationally efficient. The validity of the controller is demonstrated with simulation results.Index Terms--Chebyshev neural network, nonlinear systems, sliding mode control, twin rotor control system.