The work aims to put forward an adaptive large-capacity watermarking algorithm based on the visual regional characteristics and diamond coding, regarding the small watermark capacity of the NSCT_SVD watermarking algorithm. The NSCT transform was performed on the image, and the low frequency coefficient matrix of the NSCT domain was divided into 8 × 8 blocks, and the largest singular value in each block was selected. According to the different sensitivities of the human eye to the brightness, texture and smoothness of the image, the watermark information of different sizes was embedded in different blocks. The experiments showed that, the PSNR of the six watermarked grayscale images with 512×512 pixels was about 39, with a better imperceptibility, and the watermark capacity was about 41.41 times higher than that of the DWT_SVD watermarking algorithm, which was about 20.95 times higher than that of the NSCT_SVD watermarking algorithm. In the meantime, it had good robustness against the common median filtering, Gaussian filtering, salt and pepper noise, rotation, JEPG compression and other attacks. The proposed algorithm has increased the watermark capacity and has certain practical value on the premises of meeting the invisibility and robustness.