The objective of the study was to demonstrate the non-inferiority of low-frequency spa therapy combined with rehabilitation (Spa-rehab) versus standard spa therapy at 6 months for symptomatic knee osteoarthritis (KOA). A prospective, randomized, monocenter, non-inferiority trial with recruitment of community-based symptomatic KOA patients was performed. Standard spa therapy comprised standardized spa treatment, 6 days a week for 3 weeks, and Spa-rehab therapy comprised spa sessions, 3 days a week for 3 weeks, followed by a dedicated rehabilitation program, 3 days a week for 3 weeks. The primary endpoint was achieving at 6 months a minimal clinically important improvement (MCII) for pain on a visual analog scale and/or an MCII for function on the WOMAC index and no knee surgery (composite MCII). Secondary endpoints were composite MCII at 3 months and achieving a Patient Acceptable Symptom State (PASS) for pain and function at 3 and 6 months. Among 283 patients included, 145 were allocated to standard spa therapy and 138 to Spa-rehab therapy. We could not demonstrate the non-inferiority of Spa-rehab therapy for the primary endpoint: difference for responders - 0.08 [90% CI (- 0.18 to 0.02), p = 0.14]. However, the difference test between the groups was not significant (p = 0.18). Spa-rehab therapy was not inferior to standard spa therapy for the composite MCII at 3 months or the PASS at 3 and 6 months. Spa-rehab therapy can reasonably be proposed to patients with symptomatic KOA. This protocol may be more cost-effective than standard spa therapy and avoid absenteeism from work and accommodation costs for patients who live close to a centre.