In order to solve the problem of poor performance stability of traditional wind turbines, a data-driven approach to optimize the performance and life prediction of wind turbines is proposed. Firstly, sensors are installed to collect the operating data of wind turbines; secondly, according to the working principle and operating characteristics of wind turbines, corresponding mathematical models are established, and multi-objective power generation optimization algorithms are designed; finally, the control system structure of the turbine is clarified, and the wind speed section is divided into full wind speed sections, and multi-objective power generation optimization is implemented in the full wind speed section. The experimental results show that: the closer the smoothing factor is to 0, the smoother the output power of the wind turbine is, and the better the multi-objective power generation optimization control effect is; the closer the smoothing factor is to 1, the greater the fluctuation of the output power, and the worse the multi-objective power generation optimization control effect is.