Inertial measurement units (IMUs) are a potentially useful tool for clinicians and researchers in assessing spine movement biomechanics and neuromuscular control patterns. This study assessed the between-day reliability of the HIKOB FOX IMU in measuring local dynamic stability (LDS) and variability of trunk movements in patients with chronic low back pain (LBP). The local divergence exponent (λmax) was used to quantify LDS and the mean standard deviation (MeanSD) between cycles was used to quantify variability during 30 repetitive cycles of flexion/extension, rotation, and complex movement tasks. For λmax the average coefficient of variation (CV) was ~10% in the flexion/extension and rotation tasks, and all CV values were <20% when also including the complex task. ICC values for λmax ranged from 0.28 to 0.81. Reliability of λmax was similar between the pelvis and thorax segments (CV: ~10%, ICC: 0.48–0.78) and worse for the lumbar spine (CV: ~15%, ICC: 0.28–0.59). The CV for MeanSD was typically in the range of 20–30%, with even greater CV in the non-primary axes during each task (30–52%). Similarly, ICC values were lowest about the anterior-posterior axis in the flexion/extension task (ICC: 0.15–0.29) and largest about the longitudinal axis in the rotation task (ICC: 0.76–0.88). The moderate between-day reliability of λmax in the sagittal and transverse planes offers improvement over manual and subjective tests with poor reliability that are currently used in clinics. The minimal detectable differences presented give a threshold for change in research and rehabilitation in patients with LBP.