With the aim of finding the appropriate halftoning algorithms for watermarking of the printed image, three commonly used types of halftoning algorithm i.e. Bayer dithering, error diffusion and green noise are compared. The watermark information is embedded in the host image using discrete wavelet transform. PSNR is used to evaluate the quality of halftone images generated by the three algorithms, NC to measure the quality of extracted watermark, and the running time of halftoning to determine the computational complexity. Combined with human visual subjective, the comprehensive evaluation has been presented for the halftone image. The experimental results show that error diffusion has the best performance for wavelet watermark image.