In this study, an adaptive nonlinear sensorless controller for doubly fed induction generator (DFIG) driven by a wind turbine is proposed. The aim is to maximize the extracted power by tracking the wind turbine optimal torque-speed characteristic without the need for rotor speed measurement. This controller ensures a satisfactory tracking of both stator flux and rotor speed. It considers the detailed model of DFIG in an arbitrary (d-q) rotating frame without any simplifying assumption. An observer provides the rotor speed estimation incorporated in the control loop. To guarantee the system stability under parametric uncertainties like the aerodynamic torque and the rotor winding resistance, which harms the efficiency and the robustness of the controller, update laws are established to estimate the uncertain parameters. Lyapunov's theory is used to prove the system stability. The proposed adaptive sensorless controller validity is demonstrated by simulation in Matlab/Simulink environment. The robustness of the controller is confirmed by the comparison between the same controller with and without adaptation.