Abstract: Aiming to the problems of low embedding capacity and poor visual quality of the extracted secret images in existing generative data hiding algorithms, a generative data hiding algorithm based on multi-scale attention was proposed. First, a generator based on multi-scale attention which had dual encoder-single decoder was designed. At its encoding end, the features of the cover image and secret image were extracted independently in two branches, and fused by a multi-scale attention module at its decoding end. Skip connections were used to provide different scale detail features, thereby ensuring high-quality of stego-images. Second, self-attention modules were introduced into the extractor of the U-Net structure to weaken the deep features of the cover image, and enhance the deep features of the secret image. The skip connections were used to compensate the detail features of the secret image, so as to improve the accuracy of the extracted secret data. At the same time, the adversarial training of the multi-scale discriminator and generator could effectively improve the image quality of the stego-image. Experimental results show that the proposed algorithm can achieve an average Peak Signal-to-Noise Ratio (PSNR)/Structure Similarity Index Measure (SSIM) of 40.93 dB/0.9883 for the generated stego-images, and an average PSNR/SSIM of 30.47 dB/0.9543 for the extracted secret images under the embedding capacity of 24 bpp.