Objectives To examine whether metabolic syndrome or its individual components predict the risk of incident knee osteoarthritis (OA) in a prospective cohort study during a 32-year follow-up period. Design The cohort consisted of 6274 participants of the Mini-Finland Health Survey, who were free from knee OA and insulin-treated diabetes at baseline. Information on the baseline characteristics, including metabolic syndrome components, hypertension, elevated fasting glucose, elevated triglycerides, reduced high-density lipoprotein, and central obesity were collected during a health examination. We drew information on the incidence of clinical knee OA from the national Care Register for Health Care. Of the participants, 459 developed incident knee OA. In our full model, age, gender, body mass index, history of physical workload, smoking history, knee complaint, and previous injury of the knee were entered as potential confounding factors. Results Having metabolic syndrome at baseline was not associated with an increased risk of incident knee OA. In the full model, the hazard ratio for incident knee OA for those with metabolic syndrome was 0.76 (95% confidence interval [0.56, 1.01]). The number of metabolic syndrome components or any individual component did not predict an increased risk of knee OA. Of the components, elevated plasma fasting glucose was associated with a reduced risk of incident knee OA (hazard ratio 0.71, 95% confidence interval [0.55, 0.91]). Conclusions Our findings do not support the hypothesis that metabolic syndrome or its components increase the risk of incident knee OA. In fact, elevated fasting glucose levels seemed to predict a reduced risk.