This thesis presents two reversible data hiding methods. The first reversible data hiding method embeds data used on integer transform and histogram shifting. The inverse transform exploits the correlations among adjacent pixels in an image. We calculate the mean and difference values of the adjacent pixels. We use different block of size and overlapping two layer difference value of histogram, not only we can embed data into the image, but also data capacity is better. So the object of data hiding is arrived. The second reversible data hiding method embeds data used on prediction and histogram. We divide difference block size that pixel values slightly to embed data into the image. Experimental results know that our capacity is better payload and PSNR. After the data hidden data have been extracted, the image can recover without any distortion.