Introduction: Mindfulness-based stress reduction (MBSR) has demonstrated its effectiveness in reducing pain-related stress in adults with chronic pain. However, the implementation of MBSR needs modifications across cultures. This pilot study reports the findings of a randomized controlled trial that investigated the effects of a culturally adaptive MBSR program on self-report and neuroimaging outcomes for chronic pain adults in China. Methods: Sixty-seven participants were randomly assigned to the treatment group (n = 40) or the treatment-as-usual group (n = 27) group at a ratio of 1.5:1. Participants completed self-report measures of pain severity, pain interference, depression, perceived stress, pain catastrophizing, mindfulness, and resilience at baseline assessment (T1), post-treatment (T2), and 3-month follow-up (T3) assessments. Functional magnetic resonance imaging (fMRI) scanning was also performed at T1 and T3 assessments. Results: For the intention-to-treat sample, the results of the mixed-effect model indicated that Group × Time interaction was significant for pain catastrophizing only (F (2, 130) = 3.51, p = 0.033). Compared with the control group, those in the MBSR group reported greater reductions in pain catastrophizing at T2 (d = − 0.60), though this effect was not maintained at T3 (d = − 0.05). Additionally, the results of completer analyses found significant Group × Time interactions for pain interference (F (2, 88) = 4.40, p = 0.015) and perceived stress (F (2, 88) = 3.13, p = 0.048), but not for other measures. Finally, both groups exhibited decreased regional homogeneity (ReHo) in the frontal lobe, while increased ReHo in the cerebellum anterior lobe was unique to the MBSR group. Conclusions: The present findings suggest that the minor modified MBSR program improves certain pain-related outcomes for Chinese adults with chronic pain. Future studies with larger samples of Chinese chronic pain patients are needed to detect the small-to-moderate benefit of MBSR on fMRI and/or other objective methods.