A novel digital watermarking algorithm based on wavelet transform and singular value decomposition method was put forward, which is resistant to a variety of attacks and has a strong robustness. At first, two levels wavelet transformation was carried out on original image. Then, singular value decomposition was carried out on low frequency sub-band to get singular value from original watermark. Finally, the watermark was embed with weight method, in which the weighting coefficients were obtained from large amounts of MATLAB simulation experiments. The experimental results showed that the algorithm has strong robustness against any angle rotation attacks, salt & pepper, speckle, Gaussian and Poisson noise attacks, various filtering attacks, and JPEG compression attacks.