This study proposes a two-stage segmentation of positive cells nucleus at the outward image of an immunohistochemical staining of breast cancer. The first stage of hue filtering was done with twice of double thresholding method to retain only the areas that have the same hue value (shades of brown) with positive cells nucleus. The second step is the separation between the range of saturation intensity of positive cells nucleus and other objects that are not positive cells nucleus by using Fuzzy White Top-hat method. This two-stage processing makes the proposed method become more stable against variations of the background shades of outward appearance and image segmentation results are better. The experimental results showed that there were an increased average of approximately 2,9 times higher both on fuzzy morphology hue-based method and double thresholding hue-based method. This method is expected to gain contribution on software development that can replace pathologist/observer role in the manual counting of positive cell nucleus outward image of immunohistochemistry.