Background/Objectives To compare ocular torsion between congenital and acquired unilateral superior oblique palsy (USOP). Subjects/Methods Retrospective review was performed on medical records of 163 USOP patients. Qualitative fundus torsional states in both eyes were determined based on locations of optic disc and fovea. Disc-fovea angles (DFA) were collected in both eyes for quantitative analysis. DFAs and the proportions of qualitative fundus torsional states in paretic and non-paretic eyes were compared between USOP patients and normal controls, and between congenital and acquired USOP patients. Results This study included 90 patients with congenital USOP, 73 patients with acquired USOP, and 66 normal controls. Most control subjects showed no torsion in both eyes (93.9%), whereas 61.1% of congenital and 46.5% of acquired USOP patients showed extorsion in either eye. More patients with congenital USOP showed fundus extorsion in the non-paretic eye (24.4% versus 12.3%) or both eyes (20.0% versus 6.8%), compared with patients with acquired USOP (P = 0.007). DFAs of paretic and non-paretic eyes were larger in USOP patients than in normal controls (P < 0.001, for both congenital USOP versus control and acquired USOP versus control). DFAs of non-paretic eyes were larger in congenital USOP patients than in acquired USOP patients (10.3°versus 8.5°, P = 0.018). Conclusions Congenital USOP showed greater fundus extorsion in the non-paretic eye, compared with acquired USOP. Fundus photographs of both eyes are necessary to understand the ocular torsion in USOP patients and the variations in fundus torsion with varying USOP aetiology.