The work aims to propose a blind watermarking algorithm for anti-rotation images based on pseudo Zernike moments and DWT-SVD. First of all, the secondary wavelet transform was done on inscribed square area in the inscribed circle of the original image. SVD decomposition was made for the low-frequency sub-bands. In a certain embedding strength, the watermark was added to the singular value matrix and the watermark embedding was achieved. Finally, the phase information of the pseudo Zernike moments of the watermark image was recorded to detect the rotation angle and to correct the rotation attack in the watermark extracting process. Simulation results indicated that the algorithm had a good visibility that the PSNR value was as high as 45.12, the similarity of the extracted watermark and the original watermark was high, and the NC value was 1. The robustness of the proposed algorithm to rotation attacks was strong and had a good robustness to the common signal processing attacks. This algorithm has strong robustness for resistance against rotation attacks at any angle, and can satisfy the needs of digital products’ copyright protection.