Colouring Chinese Paintings With the increase of market demand as well as the difficulty for painters to turn a Baimiao sketch into a coloured Chinese painting in a short period of time. Therefore there is a need for an AI technique for colouring Baimiao sketches for Chinese paintings. In this paper, we focus on the image-to-image conversion problem and propose Double Discriminator Enhanced-GAN (DDE-GAN), which is improved based on the Pix2pix method.DDE-GAN consists of a Unet structure as a generator, in which a convolutional attention mechanism module is added to focus the model's learning focus on key regions. This is followed by an elaborate enhancer. The enhancer is an augmentation block based on the perceptual field model, which enhances the dephasing effect in terms of colour and detail, and is capable of generating Chinese paintings with localisation at a fine scale. We also introduce a multi-scale discriminator that guides the generator to produce pseudo-realistic images at coarse scales. Extensive experimental results on our self-developed dataset of Chinese paintings with two art styles, brush painting and landscape painting, show that the proposed DDE-GAN outperforms the state-of-the-art method CycleGAN in terms of PSNR, SSIM, MSE, PI, and subjective visual effects.