This study aims to determine if a large anterior and reduced posterior/superior joint space is highly predictable for disc displacement. From patients with temporomandibular disorders symptoms, fifty-two experimental joints and fourteen control joints were included. The cone beam computed tomography (CBCT) images were used to calculate posterior-to-anterior (P-A) and superior-to-anterior (S-A) joint space ratios, while disc position was determined using magnetic resonance imaging (MRI). One-way analysis of covariance test and receiver operating characteristics analysis were carried out. The results showed that among the 52 experimental joints, 45 were diagnosed as disc displacement and 7 as normal disc positions (N). All 14 control joints showed normal disc positions. The P-A ratio was 1.46 ± 0.21, 0.99 ± 0.23, and 0.86 ± 0.30 in the control, N, and DD groups, respectively (p < 0.001). The S-A ratio was 1.80 ± 0.27, 1.44 ± 0.33, and 1.08 ± 0.35 in the control, N, and DD groups, respectively (p < 0.001). When an altered P-A ratio and/or S-A ratio are observed on the CBCT, the diagnosis of disc displacement is quite predictable with high sensitivity and specificity.