The variations of daily renewable energy resources (RESs) such as solar and wind follow certain patterns, as well as daily load. This paper uses clustering techniques to cluster hourly load, wind speed and solar radiation by the day, and then some typical scenarios are obtained together with their probabilities of occurrence. These scenarios are applied to energy storage system (ESS) design in stand-alone microgrids, comprising diesel generators, wind power generating system and solar power generating system. The ESS minimizing the total cost over the whole project lifetime is suggested. Finally, a simple stand-alone microgrid is introduced to verify the efficiency of the proposed methods.