A sound retrieval method described in this paper enables users to easily obtain their desired sound. A sound representation experiment was conducted to study how people represent the sounds. Almost all representations were made with verbal descriptions that could be classified into “description of sound itself”, “description of sounding situation” and “description of sound impression”. The retrieval method, which adopts three keyword types, onomatopoeia, sound source, and adjective, was proposed based on the experimental results. The sound retrieval system’s efficiency was discussed based on the subjective evaluation. Users can select a convenient retrieval method and adapt it to their idea of retrieval.