Intelligent devices such as robots and drones have become the main carriers for substation inspections. However, due to the inherent noise of these devices, most of them can only be equipped with optical sensors such as visible light and infrared. This article studies the unmanned aerial vehicle voiceprint detection technology for substation equipment based on array noise reduction technology. By using array noise reduction technology to solve the problem of wind noise impact on unmanned aerial vehicles, and then combining it with equipment sound cutting processing, the sound signal of the equipment under the background noise of unmanned aerial vehicle flight is reconstructed, providing a new solution for targeted monitoring of equipment abnormalities in noisy environments.