AIMS: To evaluate the quality of methods used in randomized controlled trials (RCTs) of treatments for management of pain and dysfunction associated with temporomandibular muscle and joint disorders (TMJD) and to discuss the implications for future RCTs. METHODS: A systematic review was made of RCTs that were implemented from 1966 through March 2006, to evaluate six types of treatments for TMJD: orthopedic appliances, occlusal therapy, physical medicine modalities, pharmacologic therapy, cognitive-behavioral and psychological therapy, and temporomandibular joint surgery. A quality assessment of 210 published RCTs assessing the internal and external validity of these RCTs was conducted using the Consolidated Standards of Reporting Trials (CONSORT) criteria adapted to the methods of the studies. RESULTS: Independent assessments by raters demonstrated consistency with a mean intraclass correlation coefficient of 0.63 (95% confidence interval). The mean percent of criteria met was 58%, with only 10% of the RCTs meeting the four most important criteria. CONCLUSIONS: Much of the evidence base for TMJD treatments may be susceptible to systematic bias and most past studies should be interpreted with caution. However, a scatter plot of RCT quality versus year of publication shows improvement in RCT quality over time, suggesting that future studies may continue to improve methods that minimize bias.