Segmentation is one of the basic problems in MRI analysis. We consider the problem of simultaneously segmenting multiple MR images, which, for example, could be a series of (2D/3D) images of the same tissue scanned over time, different slices of a volume image, or images of symmetric parts. The multiple MR images to be segmented share common structure information and hence they are able to assist each other in the segmentation procedure. We propose a Bayesian co-segmentation algorithm where the shared information across images is utilized via a Markov random field prior, and a Gibbs sampler is employed for efficient posterior sampling. Because our co-segmentation algorithm pulls all the image information into consideration simultaneously, it provides more accurate and robust results than the individual segmentation, as supported by results from both simulated and real examples.