This paper investigates the capacity optimization of three different renewable-based systems for standalone households in South Australia. Solar photovoltaic (PV), wind turbine (WT) and battery energy storage (BES) are the main components for such systems. Three selected configurations of the system are: PV with battery, WT with battery, and PV-WT with battery. Optimal capacity of each component is determined through an optimization process by considering the system net present cost as an objective function. The optimization is based on the one-year hourly real data of load consumption, wind speed, solar insolation and air temperature of a remote area. Capacities of the system components are first optimized for a case of uninterruptible supply of the load consumption. It is demonstrated that such a system results in high cost and large capacity of energy storage. Two strategies are then proposed to reduce the components capacities and hence the total system cost. The first strategy is to reduce the peak demands manually. In the second strategy, loss of power supply probability is used as a reliability index to reduce the capacity of the components and the total cost substantially. Sensitivity analysis are also carried out based on the component's capital cost and daily average load consumption for the optimum configuration. Annual operational results are presented and discussed for the optimal configuration.