With the rapid development of technology, China’s emphasis on new energy is gradually increasing. The development and application of new energy can not only reduce the impact of energy crisis on the country’s industrial and technological development, but also avoid the inevitable pollution of the surrounding environment caused by the use and development of energy. Wind power generation is one of the current new energy sources, which mainly uses wind power to drive the gearbox in the generator set to rotate, thereby converting wind energy into electrical energy. However, in the process of using wind energy to generate electricity in wind turbines, their gearboxes are often subjected to significant impacts, leading to malfunctions. To avoid getting stuck in solving complex nonlinear signals, a digital information frequency screening and processing method based on the fusion of multiple algorithms is proposed. Firstly, wavelet analysis is used to denoise the sensitivity of high-frequency signals. Then, the Hilbert Huang transform is fully utilized to decompose non-stationary signals and explore their time-frequency transformation capabilities for signal feature mining, in order to identify the corresponding frequency of faults in the disturbed non-stationary signals. Fully utilize multiple algorithm fusion to effectively remove high-frequency noise unrelated to the main frequency and identify the marginal spectrum of the time-domain transformation. Avoiding the uncertainty of directly decomposing high-frequency signals, reducing the reflux components in the signal decomposition process, avoiding reflection phenomena, and completing the frequency screening and comparison of non-stationary signals for wind turbine gearbox faults. By conducting a fault experiment on a certain unit, the effectiveness of the digital information frequency screening method was verified.