This paper presents an improved BP neural network nodel enhanced with the Sparrow Search Algorithm (SSA) for more accurate power forecasting of wind turbines. The Sparrow Search Algorithm is an efficient global optimization technique inspired by the foraging behavior and predator evasion strategies of sparrows. In this study. SSA is utilized to optimize the weights and bins-es of the BP neural network, aiming to address the issue of local minima commonly encountered by traditional BP networks when processing complex wind power data. By integrating the Sparrow Search Algorithm, the improved BP neural network can more effectively learn the nonlinear relationships between historical meteorological data and power output, thereby enhancing the model’s predictive performance. The study concludes by comparing the forecasting results with those optimized by the Vulture Optimization Algorithm, demonstrating the superiority of the SSA-enhanced BP neural network in predicting the power output of wind turbine units with greater accuracy.