The existing underwater enhancement algorithms have some problems such as color distortion and bad defogging effect. Therefore, this paper proposes an underwater image enhancement algorithm based on dual attention mechanism and improved U-Net. Firstly, the color correction module is used to process the red, green and blue channels to reduce the influence of color deviation. Then, the channel attention and spatial attention are fused with the U-Net network, and the images after color correction are defogged and denoised to retain the texture details of the images and enhance the contrast. Finally, the pyramid fusion module is used to fuse the image features with different resolutions to obtain a clear visual image. The experimental results show that based on UIEBD and UFO-120 test sets, the average values of UCIQE, NIQE, SURF and information entropy are 0.608 1, 4.440 3, 31.5 and 7.649 5, respectively. The proposed algorithm is superior to other classical and novel algorithms in subjective visual quality and objective evaluation indexes. The enhanced underwater image has good defogging effect and obvious advantages in color correction, which significantly improves the visual quality of underwater images.