The aim of this study was to evaluate the psychometric characteristics of three versions of the Dutch Oral Health Impact Profile (OHIP-NL), for clinical use with temporomandibular disorder (TMD) patients. To that end, two abbreviated OHIP versions (OHIP-NL14 and OHIP-NL5) were developed by respectively selecting 14 and five items from the officially translated and culturally adapted original 49-item OHIP-NL questionnaire. A total of 245 consecutive patients, referred by their dentist to the TMD clinic of the Academic Centre for Dentistry Amsterdam (77% women; mean age ± s.d. = 41·0 ± 14·9 years), completed the Research Diagnostic Criteria for TMD (RDC/TMD) axis II questionnaire and the OHIP-NL. Reliability and validity of all three OHIP versions were compared, and their associations with four psychological axis II variables, indicating the level of impairment of patients with TMD, were examined. According to guidelines for clinical application, internal consistency scores were sufficient for OHIP-NL and OHIP-NL14, but insufficient for OHIP-NL5. Test-retest reliability (n = 64) was excellent for OHIP-NL and OHIP-NL14 and fair to good for OHIP-NL5. For all three versions, there was evidence for score validity: associations between OHIP summary scores on the one hand and validation variables and other RDC/TMD axis II variables on the other hand met the expectations and were statistically significant (P < 0·001). In conclusion, the OHIP-NL and OHIP-NL14 both performed comparatively well and better than the OHIP-NL5. When the length of the questionnaire (i.e. the time needed for its completion) is an issue, the OHIP-14 would therefore be the preferred version.