A robust adaptive backstepping control scheme is proposed for a class of pure-strict non-affine nonlinear system with state constraint and input saturation. Taylor series expansion technique is applied to the non-affine system to convert it to affine-like expression with high accuracy. The recurrent perturbation fuzzy neural networks disturbance observer (RPFNNDO) based on the projection algorithm is designed to approximate the unknown compound disturbance online. Backstepping control is used with the barrier Lyapunov function, the tanh function and the Nussbaum function to design controllers, which handles states constraints, input saturation in the system. The stability of the closed loop system is analyzed by using the Lyapunov theory. Simulation results of the unmanned aerial vehicle track control show the effectiveness of the proposed method.