Background. Osteoarthritis is the most frequent chronic joint disease which causes pain and disability of especially hip and knee. According to international guidelines and the Dutch general practitioners guidelines for non-traumatic knee symptoms, acetaminophen should be the pain medication of first choice for osteoarthritis. However, of all prescribed pain medication in general practice, 90% consists of non-steroidal anti-inflammatory drugs compared to 10% of acetaminophen. Because general practitioners may lack evidence showing a similar efficacy of acetaminophen and non-steroidal anti-inflammatory drugs, we present the design of a randomized open-label trial to investigate the efficacy of a non-steroidal anti-inflammatory drug (diclofenac) compared with acetaminophen in new consulters with knee osteoarthritis in general practice. Methods/Design. Patients aged 45 years or older consulting their general practitioner with non-traumatic knee pain, meeting the clinical American College of Rheumatology criteria, and with a pain severity score of 2 or higher (on a 0-10 scale), will be randomly allocated to either diclofenac (maximum daily dose of 150 mg) or acetaminophen (maximum daily dose of 3000 mg) for 2 weeks and, if required, an additional 1-2 weeks, with a total follow-up period of 12 weeks. The primary outcomes are knee pain measured with a daily diary, and pain and function measured with the Knee Injury and Osteoarthritis Outcome Score (KOOS) at baseline, and at 3, 6, 9, and 12-weeks follow-up. Secondary outcomes are patients' perceived recovery, quality of life, medical, patient, and productivity costs, compliance to therapy, co-interventions, and adverse reactions. Discussion. The successful completion of this trial would lead to a better understanding of which medication should be used in the treatment of primary care patients with mild knee osteoarthritis. Trial registration. Dutch trial registry NTR1485.