In this study, fuzzy input data has been considered for processing of Poisson medical image matting. Image matting is actually known as an image segmentation approach. But, fine detail information can be extracted from the background of the given image in the image matting. Although global Poisson image matting approach is applied to smoothed images, successful matting results can be obtained from medical images that may not be contain more fine detail information. In this study, fuzziness is included to the input data as a percent value and a generated result is compared with the classical Poisson image matting approach.