In view of having adaptive controllers for nonlinear systems that do not take into account external disturbances and all degrees of freedom of the system, this paper proposes a robust identification and control-based neural network method for a Twin Rotor Multivariable System (TRMS) using a recursive adaptive descendent gradient algorithm adagrad in discrete time. The neural network identification is performed online and the TRMS is controlled under a polynomial structure by pole placement. The method results obtained by MATLAB simulations are evaluated in terms of estimation and tracking error in the presence of external disturbances and sinusoidal reference signals.