The purpose of this pilot study was to use a multidimensional model to evaluate deficits in patients with low back pain (LBP) over the course of a 12-week rehabilitation program using the Swiss ball.A within-subjects, repeated-measures design based at the University exercise training clinic was used. Twenty patients with chronic nonspecific LBP (12 men, 8 women; symptom duration, 4.8 years; 38.8 ± 12.1 years old; height, 1.76 ± 0.06 m; weight, 76.15 ± 7.21 kg) participated in this study. Self-report measures were the Oswestry Disability Index, Visual Analog Scale, Medical Outcomes 12-Item Short Form Health Survey, and Self-Efficacy For Exercise Scale. Physiologic measures were electromyography measurement of feedforward muscle activation, flexion relaxation phenomenon, myoelectric fatigue, endurance capacity measured by the Sorenson test, and a modified sit-up test. Individuals performed 12 weeks of progressive exercise periodized every 4 weeks using a Swiss ball. Outcome measures were assessed at baseline, 4 weeks, 8 weeks, 12 weeks, and at a 3-month follow-up. Repeated-measures analysis for variance for time differences and regression analysis for variance in Oswestry scores were performed.The Oswestry score for self-reported disability significantly decreased over the intervention (F 4,14 = 19.456, P < .001). Significant improvements in pain and disability maintained to the 3 months of follow-up. There were significant changes in perceptions of physical and mental well-being, erector spinae fatigue, and flexion relaxation measures. Change in flexion relaxation explained 38% of the improvement in Oswestry scores at the 12-week measurement.This study showed that the Swiss ball may be successfully used in a rehabilitation context for patients with LBP. This pilot study has used a novel approach to assess improvements during a rehabilitation program, which may be used in the future to explain differences between different treatment modalities.