With the increase of pyrotechnic operations in the construction of electric power projects, the safety problems in the operation process become more and more prominent, so the intelligent detection technology of pyrotechnic operations is of great significance in the safety management. This paper investigates the application of the YOLOv9 model in the detection of pyrotechnic operations, including the detection of whether there is pyrotechnic behavior and whether the operators wear protective equipment. The design concept and advantages of the YOLOv9 model are first introduced, followed by a detailed description of the model’s architecture, training process, and optimization strategy in the detection of pyrotechnic operations. By using the self-constructed dynamic fire operation dataset for training and validation, the experimental results show that the YOLOv9 model exhibits good detection performance in different operation scenarios. The study provides effective technical support for the safety management of power engineering construction sites and has important theoretical significance and practical application value.