This paper proposes a design of a direct optimal control for a class of multi-input–multi-output (MIMO) nonlinear systems. This work focuses on the design of optimal model-free backstepping controller for a MIMO quadrotor helicopter perturbed by unknown external disturbances. The proposed method consists of using a model-free-based backstepping controller optimized by a cuckoo search algorithm. First, the overall dynamic model is decoupled into six interconnected subsystems. Then, the ideal backstepping controller with a known dynamic function is designed for each subsystem. The model-free based on backstepping control uses a new estimator approach to approximate the unknown dynamic model functions. After that, the global asymptotical stability of the closed-loop control system is proved via the Lyapunov theory. Moreover, the parameters of the proposed controller are optimized by the cuckoo search algorithm according to a cost function. The results of numerical simulations applied to the quadrotor helicopter system demonstrate the robustness and the effectiveness of the proposed control strategy.