Abstract: An image-adaptive watermarking algorithm based on wavelet transform was proposed. At first, A digital image used as watermark was scrambled. Next, the original image and scrambled watermark image was decomposed by discrete wavelet transform(DWT), and the Markov random field model was used to obtain the allowable distorted parameter of every lowest approximate coefficient, embedding position; and intensity were adjusted according to the allowable distorted parameter. At last, the lowest approximate coefficients of watermark was embedded into lowest approximate coefficients which were selected from original image.