In the most recent couple of decades, medical image processing stood out within picture preparing research fields because of its nonintrusive nature. Restorative imaging modalities, for example, MRI, CT filter, for the most part, rely upon computer imaging innovation to create or show advanced pictures of the inward organs of the human body, which causes the medicine professionals to envision the internal bits of the body. Here the proposed algorithm is thresholding different tissue type of brain MR image. Modes of the histogram represent different tissue types in brain MR image. So, this algorithm depends on the principle of finding maxima and minima using differentiation of the smoothed histogram. Using discrete differentiation, the author finds the multiple thresholds of brain MR image by selecting proper location of minima. The algorithm can be used as an initial segmentation of different tissue types of brain MR image for further accurate detection of the regions.