Trismus after head and neck cancer treatment may severely limit mandibular functioning. Interventions aimed at reducing trismus can only be evaluated when the amount of variation associated with these measurements is known. The aim of this study was to analyse the variation in mouth-opening measurements in patients treated for head and neck cancer, with and without trismus. Maximal mouth opening was measured in 120 patients in two sessions of three repeated measurements by one observer. To analyse the influence of interobserver variation on mouth-opening measurements a subgroup of 30 patients was measured by a second observer. The standard deviation of the six measurements per patient was used as the variation in measurements of maximal mouth opening. No significant difference was found in maximal mouth opening in patients with (n=33) or without (n=87) trismus. The interobserver intraclass correlation coefficient (ICC) was 0.98. Intraobserver ICC and intersession ICC reliabilities both were 0.99. The variation in the mean values of the three measurements was only slightly smaller than the variation of the single measurements. Variation in maximal mouth opening in patients with trismus does not differ from variation in maximal mouth opening in patients without trismus. Interobserver variation is limited.