This study aimed to compare the long-term effects of flexion- and extension-based lumbar exercises on chronic axial low back pain (LBP). This was a 1-year follow-up of a prospective, assessor-blind, randomized controlled trial. Patients with axial LBP (intensity ≥ 5/10) for > 6 months allocated to the flexion or extension exercise group. Patients underwent four sessions of a supervised treatment program and were required to perform their assigned exercises daily at home. Clinical outcomes were obtained at baseline, 1, 3, 6 months, and 1-year. A total of 56 patients (age, 54.3 years) were included, with 27 and 29 in the flexion and extension groups, respectively. Baseline pain and functional scales were similar between both groups. The mean (± standard deviation) baseline average back pain was 6.00 ± 1.00 and 5.83 ± 1.20 in the flexion and extension groups, respectively. At 1-year, the average pain was 3.78 ± 1.40 and 2.26 ± 2.62 (mean between-group difference, 1.52; 95% confidence interval 0.56–2.47; p = 0.002), favoring extension exercise. The extension group tended to have more improvements in current pain, least pain, and pain interference than the flexion group at 1-year. However, there was no group difference in worst pain and functional scales. In this controlled trial involving patients with chronic axial LBP, extension-based lumbar exercise was more effective in reducing pain than flexion-based exercises at 1-year, advocating lumbar extension movement pattern as a component for therapeutic exercise for chronic LBP.Clinical Trial Registration No.: NCT02938689 (Registered on www.clinicaltrial.gov; first registration date was 19/10/2016).