Recent studies have shown higher brain functions like motor control and cognition are coupled with the breathing cycle. Here we investigated if breathing is involved when participants perform a time estimation task. In the talk, I will present evidence supporting an involvement of breathing in time estimation. Participants estimated the end point of a short duration (e.g. 500 ms) by giving a keypress. The start point of the duration was either marked by a visual stimulus (passive condition) or by a keypress from the participants (active condition). The estimated duration was significantly shorter in the active condition than in the passive condition (i.e. the temporal binding effect). Importantly, significant differences in the breathing pattern was also found between the two conditions. In general, participants tended to start exhalation after the duration start point. However, the exhalation started earlier in the active condition than in the passive condition, which is consistent with the earlier indication of the duration end point in the active condition than in the passive condition. Furthermore, the time of exhalation predicted the results of time estimation at the individual level. The current study highlights breathing as an internal cue for time estimation, at least when the estimated time is short (e.g. within 1 second).