Abstract: Considering the problems of waste of manpower and poor supervision effect in the traditional methods, this paper studies a safety helmet wear detection method for coal miners based on the helmet target image detection method, and uses the newer YOLOv8 algorithm to realize the helmet target detection. Firstly, a helmet target detection model based on YOLOv8 algorithm is constructed. The experimental environment of helmet target detection based on Pytorch depth model and CUDA module is built in Ubuntu environment. 4000 images of actual manufacturers in coal mines are collected and made into training and verification data sets, and the helmet target detection model based on YOLOv8 algorithm is trained. The final training model has a detection accuracy of 94.4%, which meets the detection needs of helmet wearing in coal mines.