Based on the print-scan invariant, a multi-contour pixel flipped watermarking algorithm for text images was proposed, which contains three measures as follows. First, a watermark compression preprocessing method was proposed. By compressing watermarking information, the number of times of watermark embedding is increased under the condition of a certain total embedding capacity, so as to improve the robustness of the watermark. Second, a multi-contour pixel flipping strategy was proposed, which can reduce the possibility of false embedding of the watermark, and further improve watermark robustness. Finally, characterized by print-scan invariant, combined with discrete cosine transform and multi-contour pixel flipping strategy, an optimized dual-domain combined watermark embedding and extraction algorithm was proposed to improve the ability of the watermark to resist print-scan attack while ensuring the invisibility of the watermark. Experiments show that the proposed algorithm can improve the robustness of the watermark against print-scan attack while ensuring the invisibility of the watermark, and it can also resist common image attacks.