Wind farms must comply with requirements of grid-connected standards; the power fluctuations limit is one of the most challenging requirements. A promising solution to mitigate the impact of wind power fluctuations is to employ battery energy storage systems (BESSs) at the grid connection point of the wind farm. This paper proposes a nonlinear least squares method to smooth wind power fluctuations by using BESSs. The proposed method can minimize the sum of output power squared errors and reduce the size of the BESSs required to smooth the wind power fluctuations. Simulation results show that the proposed control method returns better results in smoothing the wind power fluctuations compared to linear least squares and moving average method.