To reduce image distortion and enhance the steganalysis resistance capability of conventional steganography algorithms at high-capacity data embedding, large-capacity residual plane steganography based on multiple adversarial networks is presented in this paper. Instead of directly modifying pixels of an original carrier image, the proposed approach embeds secret messages into the residual plane of the two most similar channels within the carrier media. Thereby, enabling the distortion of the carrier image imperceptible even at a same-sized image has been embedded. At the same time, multiple steganalysis networks are parallelly settled to enhance the steganalysis resistance capability of the stego image. Moreover, the Lion optimizer is employed in the steganography scheme for the first time to ensure rapid and stable convergence of the network. Additionally, multiple loss functions are organically combined to further enhance both the visual quality and steganalysis resistance capability of the resulting stego image. Experimental results demonstrate that the proposed scheme achieves an average PSNR of 38.91dB on stego images, and clearly outperforms its counterparts.