The developments research in the cluster analysis using the fuzzy method. The fuzzy method allocates to each group with membership value located at interval [0, 1], showing the magnitude of the possibility of an object being a member into a particular group. Outlier in data very important known before grouping, because affect the final result. Grouping by using the mean value as the center of the group will be more sensitive than using the median value, so this research applies fuzzy c-means and fuzzy c-medoid method to the grouping of villages in Sorong Regency Year 2016 based on the underdevelopment status and examine the goodness of both methods. There are 23.2% of villages that do not change when done grouping with both methods. Overall average distance of group center object and varians in the resulting group the two methods are the same, the varins between groups of fuzzy c-means is greater than the fuzzy c-medoid method.