Little is known about self-selected speed and fast walking in people with acute low back pain. This study aimed to investigate (1) the strategies that people with acute low back pain use to change from self-selected speed to fast walking and (2) the effect of a period of treadmill walking on level of back pain. Eight participants with acute low back pain and eight matched control participants were evaluated during self-selected speed and fast walking on a treadmill. The eight participants with back pain were retested 6 weeks later when pain had resolved. Measurements were taken of (1) three-dimensional angular movements of the pelvis and lumbar spine using a videoanalysis system, (2) the timing and distance parameters of walking, and (3) pain levels as measured by a visual analogue scale. We found that to walk faster, those with acute low back pain increased stride length and the frontal plane movements of pelvic list and lumbar lateral flexion (pelvis) to a greater extent than when symptoms had resolved. We also found that 10 min of treadmill walking at self-selected speed led to a reduction in the level of back pain and that there was a high degree of negative correlation between level of back pain and stride length. An additional 5 min of fast walking did not lead to any further changes in level of back pain. These findings support clinical recommendations that the moderate physical activity of walking may be beneficial in the management of people with acute low back pain. To walk faster, people with acute low back pain may utilise strategies that had been limited at self-selected speed, without any increase in pain.