Based on the advantages of wavelet transform and singular value decomposition, a new method was put forward for watermark embedding and extraction of color images, in order to improve the robustness and safety of watermark. A novel digital watermark algorithm based on singular value decomposition and wavelet transform method was used for color space conversion of the color host image from RGB to Lab. Two levels wavelet transformation was carried out on the L components of the color host image in the standard color space, and singular value decomposition was carried out on low frequency sub-band and the original quick response code. At last, the watermark was embedded with an appropriate embedding strength coefficient obtained on the basis of human visual system. The experiments showed that the proposed algorithm had strong robustness against Gaussian, salt & pepper, speckle and Poisson noise attacks, JPEG compression attacks, small-angle rotation attacks, compression attacks and contrast enhancement attacks. In conclusion, the watermarking algorithm was able to resist common watermark attacks and could meet the requirements of digital anti-counterfeiting and copyright protection.