Temporomandibular disorders (TMD) diagnoses can be viewed as the most useful clinical summary for classifying subtypes of TMD. The Research Diagnostic Criteria for TMD (RDC/TMD) is the most widely used TMD diagnostic system for conducting clinical research. It has been translated into 18 languages and is used by a consortium of 45 RDC/TMD-based international researchers. While reliability of RDC/TMD signs and symptoms of TMD has been amply reported, the reliability of RDC/TMD diagnoses has not. The aim of the study was to determine the reliability of clinical TMD diagnoses using standardized methods and operational definitions contained in the Research Diagnostic Criteria for Temporomandibular Disorders (RDC/TMD). Data came from reliability assessment trials conducted at 10 international clinical centers, involving 30 clinical examiners assessing 230 subjects. Intraclass correlation coefficients (ICC) were calculated to characterize the reliability. The reliability of the diagnoses was fair to good. Median ICCs for the diagnoses myofascial pain with and without limited opening were 0.51 and 0.60, respectively. Median ICC for arthralgia was 0.47 and 0.61 for disc displacement with reduction. RDC/TMD diagnoses of disc displacement without reduction, osteoarthritis and osteoarthrosis were not prevalent enough to calculate ICC's, but percent agreement was always >95%. The reliability of diagnostic classification improved when diagnoses were grouped into pain versus non-pain diagnoses (ICC=0.72) and for detecting any diagnosis versus no diagnosis (ICC=0.78). In clinical decision-making and research, arriving at a reliable diagnosis is critical in establishing a clinical condition and a rational approach to treatment. The RDC/TMD demonstrates sufficiently high reliability for the most common TMD diagnoses, supporting its use in clinical research and decision making.