Introduction Human diaphragm muscle exhibits similar force-velocity-power characteristics as limb locomotor muscles when assessed in situ and in vitro. In vivo, however, only force – in the form of transdiaphragmatic pressure (Pdi) – can be accurately quantified. Thus, unanswered questions remain regarding in vivo power output of the human diaphragm. Subcostal ultrasonography offers a direct measure of diaphragm displacement velocity that may be used for the estimation of in vivo power output. Our aim was to establish the feasibility and within-day reliability of diaphragm displacement velocity and estimated power output in response to volitional and non-volitional perturbations. Methods We studied diaphragm displacement velocity in 10 healthy adults (5 women; mean ± SD age 22 ± 2 y) in response to maximal inspiratory sniffs and unilateral magnetic stimulation of the right phrenic nerve. Phrenic nerve stimulation was performed as five single (1 Hz) and paired (10, 50 and 100 Hz) twitches, immediately followed by five sniffs. After a standardised rest of 20 min, the protocol was repeated. A low-frequency (1.5-4.0 MHz), phased array ultrasound probe was positioned subcostally on the right mid-clavicular line. Using anatomic M-mode (AM-mode), displacement velocity was calculated as diaphragm displacement divided by displacement time. Diaphragm power output was calculated as the product of Pdi swing and displacement velocity. Within-day reliability was assessed with intra-class correlation for absolute agreement (k = 3; 2-way mixed effects) with 95% confidence interval (CI). Results Diaphragm displacement velocity was measured successfully in 95% (range 83-100%) of ultrasound images. Displacement velocity ranged from 9.4 ± 4.5 (sniffs) to 21.9 ± 6.7 (100 Hz) cm s−1 (mean ± SD; with a strong correlation with rate of pressure development (r = 0.985, p < 0.001).The within-day reliability was moderate to excellent for all stimulation frequencies; ranging from 0.808 [95% CI 0.325-0.958, p = 0.008] for sniffs to 0.943 [0.746-0.988, p < 0.001] for 100 Hz paired twitch. Estimated diaphragm power output ranged from 130 ± 59 at 10 Hz to 916 ± 444 cmH2O cm−1s−1 for sniffs, and demonstrated a within-day reliability ranging from 0.674 [0.069-0.943, p = 0.002] for 100 Hz paired twitch to 0.889 [0.475-0.980, p = 0.002] for sniffs. Conclusion Ultrasound-derived diaphragm displacement velocity is quantifiable in response to volitional and evoked diaphragm contractions, and demonstrates a moderate to excellent within-day reliability. The measure may be used in the estimation of in vivo diaphragm power output and furthermore offers a potential new insight into the in vivo force-velocity-power characteristics of the human diaphragm muscle.