Abstract: It is difficult to balance the invisibility and the robustness of watermarks in digital watermarking algorithms. Therefore, a digital watermarking algorithm combining DNA-DWT-FRFT is proposed. The watermark image is scrambled and encrypted using DNA coding, and a watermark ciphertext sequence is obtained. Divide the carrier image into 8× 8 Block, then perform the discrete wavelet transform (DWT) on each sub-block; select the coefficients of the low-frequency sub-band and calculate their eigenvectors; finally, fractional Fourier transform (FRFT) is used to transform the eigenvectors from time domain representation to frequency domain representation. An improved DWT transform is proposed. It defines the random factors F₁ and F₂, updated according to the watermark ciphertext sequence. Then determine the watermark embedding position and update the eigenvectors of the carrier image by F₁ and F₂. Using the inverse FRFT transform and the inverse DWT transform, the eigenvectors of the carrier image containing the watermark represented are recovered back to the time domain from the frequency domain. The watermark embedding is completed. The experimental results show that the watermarking algorithm has high invisibility and strong robustness in common attack tests, such as noise attacks, rotation attacks, cutting attacks, JPEG compression attacks, etc.