Abstract: In traditional tracking systems, the mobiles report their location to server periodically, which will result in high packet loss rate and rapid energy depletion as the number of mobiles increase. Actually, in practical tracking applications, it is observed that nodes are often close to others. Hence, it is conceived to pick out some nodes to report periodically as delegates for their adjacent ones. By exploiting this thought, an effective location updating mechanism (LUM) is proposed for tracking systems in wireless sensor network. In this method, mobiles update location information through two kinds of delegates: remote and nearby delegates. Remote delegates are infrastructure nodes appointed by server. Nearby delegates are heads of clusters constructed according to the RSSI (received signal strength indicator) values. In LUM, only delegates report location periodically instead of each mobile node. Therefore, LUM can save energy greatly through reducing the message complexity. However, in practical environment, signal fluctuations will affect the process of LUM. In order to solve this problem, the parameterized flip-flop filter and strap thresholds methods are developed to smooth and stabilize the RSSI values respectively. To demonstrate the performance of LUM, a prototype system with 38 Micaz nodes are deployed. The results show that LUM outperforms traditional approaches by at least 45% less message transmission and 48% fewer energy depletion on average.