BACKGROUND: Therapeutic nuclear magnetic resonance therapy (MRT) works based on the electromagnetic fields. AIM: To investigate efficacy of MRT in knee osteoarthritis (OA). DESIGN: Prospective, randomized, double-blind, placebo controlled trial. SETTING: Outpatient clinic, university hospital. POPULATION: Patients who had mild to moderate knee OA at a single knee joint and between 30-75-years-old were randomized by blinded chip cards (1:1). METHODS: The treatment group received ten sessions of one hour daily MRT, controls received placebo MRT. All patients underwent clinical examination at baseline, after 2 weeks, and 12 weeks. Imaging included blindly assessed ultrasonography and magnetic resonance (MR) of the knee. RESULTS: Ninety-seven patients completed the study. Both groups improved significantly but the average change from baseline in outcome parameters was similar in MRT group (on VAS-pain,-2.6; WOMAC-pain, -2.09; WOMAC-stiffness, -1.81; WOMAC-physical, -1.96) compared to placebo after two weeks (VAS-pain,-1.6; WOMAC-pain, -1.91; WOMAC-stiffness, -1.27; WOMAC-physical, -1.54). Also changes were quite similar at the 12th week after the treatment. SF-36 components at 12th week improved but changes were not significant. Imaging arm also failed to show significant differences between groups in terms of cartilage thickness on US and MR scores. No adverse events were recorded. CONCLUSIONS: MRT is safe, but not superior to placebo in terms of improvement in clinical or imaging parameters after a 10-day course of treatment in mild to moderate knee OA. CLINICAL REHABILITATION IMPACT: The present study does not promote use of a 10-day course of MRT in mild to moderate knee OA.