<p>Automobile wind noise is one of the main noise sources of new energy vehicles. Active noise control (ANC) technology has mature solutions in the fields of engine order noise and road noise control. Therefore, this paper explores the methods of active control of wind noise. Firstly, simulation analysis and bench verification were carried out for the feedforward and feedback systems. Secondly, a modified FxLMS-based ANC system using secondary path equalization method was proposed. With colored noise signals and real vehicle wind noise signals as inputs, respectively, a comparative analysis was conducted with the conventional FxLMS system, subband system and frequency-domain system. Subsequently, the principal component analysis method was adopted, combined with the fisher information matrix method and the genetic algorithm to select the reference signal, respectively. Through the simulation of the wind noise of real vehicles, it was found that when the signal selected by the genetic algorithm was used as the input, the system noise reduction was 5.11&#xa0;dB(A), which was higher than that of the fisher information matrix method. Finally, real vehicle experiments were conducted based on the proposed modified system and the reference signals obtained by the two methods. The results show that when the reference signal set selected by the genetic algorithm is used as the input, the noise reduction amount of the system is higher than that of the fisher information matrix method, both reaching more than 1.55&#xa0;dB(A).</p>