White matter hyperintensities (WMH) are areas of lost cells found in the white matter of the brain presenting hyperintensities. WMH segmentation is the initial step to diagnose many brain diseases. Here, we propose an automated method of WMH segmentation designed to deal with brain magnetic resonance imaging (MRI) in polar coordinate system. Moreover, the pattern of clustering used in segmentation is adjusted to achieve the desired cluster properly. Experimental results on cross-sectional images of fluid-attenuated inversion recovery (FLAIR) datasets reveal better performance of the proposed method with simplicity and robustness. The method provides good average accuracies (0.808, 0.803, 0.981 of dice similarity coefficient (DSC), sensitivity (sens) and specificity (spec) respectively) of WHH segmentation in comparison with other method.