The paper presents the issue of power control law synthesis, in the case of a large wind system that operates under full-load regime, based on dynamic properties details in frequency domain. Solving this problem involves two phases: the establishment of a linearized model as faithfully as possible in various operating points of the full-load region, and synthesis of the power controller, considered with classic structure, taking into account frequency particularities of the obtained linearized model. Obtained linear model of the controlled process is of order 16 and encloses subsystems for tower fore-aft oscillations damping, and for drive-train torsion oscillations damping. The designed controller contains a PI component and a lag compensator for dynamic correction at high frequencies. It is known that the main features of wind system dynamics generated by the interaction of wind-tower-blade ensemble cause a gap in the gain characteristic of the model and complex conjugate zeros, which can move between right and left half-planes, depending on the average wind speed value. Consequently, for control law synthesis an interactive frequency solution is adopted. This is "transparent" in relation to particularities induced by wind-tower-blade interaction. This solution allows evaluation of the extent to which control law is affected by the subsystem for tower oscillations damping. Given the strong dependence between the model and the mean wind speed value, a gain scheduling control law is designed. At different values of average wind speed, controller synthesis is performed in two ways: one applying the desired values to the stability reserve, and the other requiring the minimization of a performance criterion, aimed at reducing mechanical stress, while the controller parameters are kept in an area that ensures admissible values to the stability reserve. This second way allows obtaining a solution for designing a simple classic controller, dedicated to large wind turbines, to which may be imposed requirements for mechanical fatigue mitigation. The proposed solutions are backed by numerical results obtained through numerical simulation. (C) 2014 Elsevier Ltd. All rights reserved.