Background: Tai Chi (TC), as one of mild to moderate exercise therapies speci cally recommended by multiple clinical guidelines, is a viable option for chronic non-speci c low back pain (CNLBP) treatment. Nevertheless, limited studies focused on the effect of different weekly frequencies of TC in elders with CNLBP. This study aims to compare the differences of TC with different weekly frequencies in elders with CNLBP on the premise of proving its effectiveness, and to identify whether mindfulness mediates the effect of TC on treatment outcomes. Methods: Totally 148 senior citizens aged 60 years old or above with CNLBP (as de ned by the National Institutes of Health criteria) will be recruited in this single-centre, randomised, single-blinded, parallel controlled trial and conducted in the Sports Centre of Changsha First Social Welfare Institute, Hunan Province, China. Participants will be randomly divided into four groups: three different weekly frequencies Chen-style TC groups (1, 3, and 5 sessions/week, 60min/session, sustain for 12weeks on the basis of weekly health educational lecture), whereas the control group will only undergoweekly healthy educational lecture. Three months of follow-up visits will be conducted. The primary outcome is pain intensity, which will be measured at baseline, after every four weeksof intervention, and after the followup period via Visual Analogue Scale (VAS). Secondary outcomes include Beck Depression Inventory-II (BDI-II), Pain Catastrophising Scale (PCS) and Five Facet Mindfulness Questionnaire (FFMQ), Oswestry Disability Index (ODI), and Short Form-36 (SF-36), which will be measured before and after intervention and after the follow-up period. The intention-to-treat and per-protocol principles will be used to analyse primary and secondary outcomes with a setting at α = 0.05 as statistical signi cance.Discussion: This comprehensive and detailed protocol will be the rst trial to compare the effectiveness of different weekly frequencies of TC in elders with CNLBP. The outcomes may provide valuable data about the choice of ideal number of sessions to further normalise the application of exercise for clinicians.