Large-scale wind power connected to the grid efficiently reduces fossil fuel consumption, but extremely decreases grid inertial and increases frequency regulation pressure on the grid. Therefore, various wind farm-based frequency regulation technologies have been investigated in recent decades. Adaptive inertial droop control of wind turbines was considered as one of the most effective methods to enhance the inertia of the grid, because it can solve the decoupling problem between the power of wind farms and power system frequency. However, the present approaches mainly pay attention to the first frequency drop (FFD) or ignore the influence of control parameters. Hence, this paper proposes a black widow optimization algorithm (BWOA)-based step start-up adaptive inertial droop controller to smooth frequency fluctuation as well as alleviate FFD, the secondary frequency drop (SFD), and the third frequency drop (TFD). Besides, BWOA is employed to extract the best parameters of the designed controller under a 150-MW load increase. Then, the extracted parameters are used in three other load variation events to evaluate the performance of the proposed method in MATLAB/Simulink. Simulation results indicate that BWOA acquires satisfactory performances on various designed load variations. Compared with the trial-and-error method, FFD and TFD with BWOA under load increase are decreased by 10.9% and 12.8% at most, respectively.