Abstract: In order to improve the real-time and accuracy of underwater target detection, an underwater target detection network model FAttention-YOLOv5 based on F-CBAM attention mechanism is proposed. The model uses the single-stage target detection network model YOLOv5 as the basic model, embeds the proposed F-CBAM attention mechanism in the model, references the FReLU activation function in the CBAM structure, and captures the complex feature distribution through two-dimensional space in the activation function stage, so as to realize the spatial information modeling ability at the pixel level and improve the accuracy of the model; The channel attention mechanism and spatial attention mechanism in F-CBAM are used to improve the channel weight of the target object and expand the receptive field of the target to the original image, so as to improve the feature learning ability of the target detection model; The recursive network feature pyramid is fused in the FAttention-YOLOv5 model, and the network can fully learn the image features of different scales through feature recursion, so as to improve the detection accuracy of small targets; Finally, the loss function of the improved model is optimized to avoid the disappearance or explosion of the gradient of the new model. The experimental results show that the designed underwater target detection model FAttention-YOLOv5 can improve the feature extraction ability of the model, effectively improve the accuracy of underwater target detection, and provide a new solution and technical assistance for marine biological capture.