Wireless sensor networks, which consist of numerous sensor nodes deployed in a specific environment, have been the focus of extensive research for monitoring physical and environmental activities. The crucial component of a sensor node is its battery power supply. As these batteries cannot be easily recharged in complex environments, achieving optimal energy consumption is a significant challenge for applications requiring a long network lifetime. Various methods have been proposed to address this challenge. One of most commonly used methods, is clustering of nodes and aggregation their data, as well as hierarchical transmission of sensed data to the base station. In this study, three types of network topologies—1) one-level, 2) two-level with maintaining dependence on the first layer cluster, and 3) two-level without dependence on the first layer cluster —were simulated using three methods of cluster head selection: 1) random, 2) sorted, selecting nodes with higher energy , and 3) weight-oriented, based on the remaining energy of the nodes and the distance with other selected cluster heads. Selecting the optimal cluster head is an NP-complete algorithm, which entails high computational complexity for implementation. The simulation results demonstrate that selecting the cluster head based on the weight-oriented method in the assumed network types increases the network lifetime, reduces energy consumption, and prolongs the stability period, measured from network initiation to the first node failure.