A moving image watermarking method, which forms a global masking map by combining frequency masking, spatial masking and motion masking values together and inserts a watermark according to the formed global masking map through the use of a human visual system is provided. In the moving image watermarking method, a watermark value is obtained by exclusive-ORing a random key value and a binary value of a logo image. A plurality of masking operations are separately performed. A global masking value is obtained through the separate masking operations. A watermarked frame value is obtained by adding a watermark value weighted by the global masking value and a control variable to an original frame value.