Due to increasing awareness on importance of personal health and recent developments in communication technology, the use of wireless sensor network in healthcare industry increases steadily. Patient comfort and prolonged battery lifetime are the key requirements in battery-powered medical devices irrespective of the methods being applied. Wearable wireless sensor nodes, that are attached to an individual in order to monitor the physiological parameters, are ideal candidates for power conservation. In this thesis, a configurable power consumption model is developed to estimate average power consumption of the ACcelerometer-based REspiration MOnitoring (ACREMO) system. The model enables the designers to predict the power consumption of each of the hardware components in the system. Various system parameters that impact the average power are identified during the modeling process. The model has estimated that the lifetime of the ACREMO system is 190 hours when driven by a 310 mAh Lithium-ion battery. A new type of respiratory monitoring algorithm is implemented on which current measurements are carried out to validate the model. The model incurs a mean error of -5.2%. Different power optimization techniques were investigated to prolong the battery lifetime of the ACREMO system and a number of proposals are made. The replacement of the accelerometer and the operational amplifiers with their low-power counterparts would save up to 41% and 7% of the system power respectively. A dynamic power management policy is presented where a set of system states are identified and their usability is defined. The proposed heterogeneous multiprocessor system showed 16% increase in the device lifetime. Reducing the CPU clock frequency and the sampling frequency resulted in 9% and 15% reduction of average power respectively. Finally, a context-aware power saving strategy is proposed that can save up to 40% of overall power by forcing the system to sleep for a predefined amount of time when a motion artifact is detected on the accelerometer signal. By carefully selecting and combining the proposed techniques, the lifetime of the ACREMO device can be prolonged from the current estimate of 7.9 days to 14 days or more with few compromises such as the accelerometer noise, hardware cost, system complexity etc.