Abstract: To accurately identify the safety hazards existing in the construction site of hydropower projects in real time, this paper introduces the channel attention mechanism and spatial attention mechanism, improves the YOLOv8 algorithm, and proposes an automatic identification method of hydropower engineering construction safety hazards based on dual attention mechanism. Firstly, based on the YOLOv8 network framework, a channel attention mechanism is constructed to adaptively highlight key features, dynamically strengthen the expression of image features of hidden danger areas, and suppress the influence of background noise. Secondly, a spatial attention mechanism is built to weight important regions, reduce background interference, optimize feature fusion, adaptively adjust attention, enhance local detail capture and positioning accuracy, improve multi-scale target detection ability, and enhance the spatial feature representation ability of the model. Finally, combined with the actual project, the accuracy and reliability of the model were verified. The results show that the proposed method can deal with the interference of hydropower project construction site through the attention mechanism, and the accuracy of construction safety hazards identification is up to 86.2%, which is better than the existing safety hazards identification model, and provides technical support for the dynamic management and precise prevention and control of hydropower engineering construction safety hazards.