A wireless sensor network (WSN) consists of spatially distributed autonomous sensors to monitor physical or environmental conditions and to cooperatively pass their data through the network to a Base Station. Clustering is a critical task in Wireless Sensor Networks for energy efficiency and network stability. Clustering through Central Processing Unit in wireless sensor networks is well known and in use for a long time. Presently clustering through distributed methods is being developed for dealing with the issues like network lifetime and energy. In our work, we implemented both centralized and distributed k-means clustering algorithm in network simulator. k-means is a prototype based algorithm that alternates between two major steps, assigning observations to clusters and computing cluster centers until a stopping criterion is satisfied. Simulation results are obtained and compared which show that distributed clustering is efficient than centralized clustering.