As people pay more attention to their health issues, different types of human activity monitoring systems are emerging in the market. Many researchers have proposed various accelerometer sensor- based mobility monitoring systems. However, the energy efficiency of wearable activity monitoring systems has not been well studied. In this paper, we develop and test an application-level solution for achieving energy savings in a human daily activity monitoring system using a wearable wireless sensor. All functionalities including data processing, activity classification, wireless communication, and storage of classified activities are implemented in a single sensor without degrading the classification accuracy of the activities. Based on the experimental protocol with five major physical activities, the system achieves an average of 98 percent accuracy in classifying these daily activities with significant energy savings.