We propose a system that controls a nap environment considering respiration rates and music tempo by using a sensor agent robot. The proposed system consists of two sub-systems. The first sub-system measures respiration rates using optical flow. We conducted preparatory experiments to verify the accuracy of this sub-system. The experimental results showed that this sub-system can measure the respiration rates accurately despite several positional relationships. It was also shown that the accuracy could be affected by clothes, movements and light. The second sub-system we constructed was the music play sub-system that chooses music with the certain tempo corresponding to the respiration rates measured by the first sub-system. We conducted verification experiments to verify the effectiveness of this music play sub-system. The experimental results showed the effectiveness of varying music tempo based on the respiration rates in taking a nap. We also demonstrated this system in a real environment; a subject entered into the room being followed by ebioNα. When the subject was considered sleeping, ebioNα started measuring respiration rates, controlling music based on the respiration rates. As a result, we showed that this system could be realized. As a next step, we would like to improve this system to a nap environment control system to be used in offices. To realize this, we need to update the first sub-system measuring respiration rates by removing disturbances. We also need to upgrade music play sub-system considering the numbers of tunes, the kinds of music and time to change music.