Background: Exercise has beneficial effects on muscle and motor function after spinal cord injury (SCI). Little is known regarding effects of prolonged intense exercise (IE) in humans with chronic SCI. Design: Prospective, non-randomized, controlled observational study. The intervention was either a multimodal IE program (n ¼ 21) or a control (CTL) intervention consisting of self-regulated exercise (n ¼ 8).Objective: Measure sensorimotor function over 6 months in relation to an IE program. Setting: Single outpatient center. Subjects: Subjects with chronic SCI (n ¼ 29 total), mainly ASIA Impairment Scale A and B, injury levels C4-T11. Results: Baseline neurological assessments (for example, ASIA motor score, 39±3 vs 42±5, IE vs CTL, P40.5, mean ± s.e.m.) did not differ between the two groups. During the 6 months, IE subjects averaged 7.3±0.7 h per week exercise, not significantly different from CTL subjects (5.2±1.3 h per week, P40.1). However, after 6 months, IE subjects showed significantly greater motor gains than CTL subjects in the main outcome measure, ASIA motor score (change of 4.8 ± 1.0 vs À0.1 ± 0.5 points, P ¼ 0.0001). The main outcome measure was calculated by ASIA motor score. These IE subject ASIA motor gains correlated with number of exercise hours per week (r ¼ 0.53, Po0.02), and with type of specific IE components, particularly load bearing. Conclusions: Multimodal IE can significantly improve motor function in subjects with chronic SCI. An organized program may provide greater motor benefits than a self-regulated program; load bearing might be of particular value. IE might have therapeutic value in chronic SCI, and as an adjunct to other restorative therapies.