Swimming speed is a function of the propulsion generated from arm strokes and leg kicks. Kicking is partially obscured underwater, making the kinematics of the kick difficult to analyse. In this study, we quantified 100-m freestyle kick-count and kick-rate patterns for 14 Paralympic swimmers using inertial-sensor technology. Swimmers took 145 +/- 39 kicks (mean +/- s) for swimming trials and 254 +/- 74 kicks for kicking-only trials. Kick rate was 124.9 +/- 20.3 kicks . min(-1) for swimming trials and 129.6 +/- 14.0 kicks . min(-1) for kicking-only trials. There were no marked differences in kick count among 25-m segments in the swimming trials. There was a substantial increase of 10.6%[90% confidence interval (90%CI): 7.3 to 14.0%] in the number of kicks in the kicking-only trials by the fourth 25-m segment. There was a substantial decrease in kick rate by the third 25-m segment for swimming (-12.0%; 90%CI: -12.8 to -11.1%) and kicking-only (-7.3%; 90%CI: -8.6 to -5.9%) trials. The relationship between swimming and kicking-only kick rates was r = 0.67 (0.55 to 0.76; P < 0.001). The temporal patterns of the kick in kicking only differed from those in swimming; increases in kick rate can improve freestyle swimming performance.