This paper proposes a realistic principal Gaussian Process model for the wind turbine, with a kernel derived. The model is trained and validated on a data-driven practical wind turbine model to predict optimal operational curves for the optimal operating conditions of wind turbine generators. The proposed technique indicates the critical parameters of the wind behaviour from historical wind speed data to keep maximum power. This technique is used to avoid periodic fluctuations in the output power of the wind turbine generators by design an efficient control system. The frequent changes in the power output come from tower shadow, wind shear, gust, and turbulence in the wind speed. The operational curves of the critical parameters included the profiles of power, blades pitch angle, and turbine rotor speed. The power improvement performance of the turbine can use the relationship between critical parameters ahead of time to predict wind behaviour.