The objective of this study was to develop and test a quantifiable method of temporomandibular (TM) joint space analysis (JSA) unaffected by head rotation. Sixteen female TM joints were imaged using multidirectional axially corrected tomography. Lateral, central, and medial image slices were obtained from each joint, traced three times each, and triplicated for drafting. Tracings and draftings were randomized, blinded, and measured manually with digital calipers. Simulated head rotations of +/-10 degrees were applied and the process repeated. Multiple comparisons were done within and between patients by analysis of variance. Based on this study, we found there was no difference for tracing, drafting, and measurement error within and between patients. There were no significant differences between -10, 0, and +10 degrees of head rotation. Our conclusion is that this method for TM-JSA has a high level of reliability as determined by analysis of tracing, drafting, and measurement error for each joint space position and simulated rotation.