Ultrasound (US) has gained popularity in the evaluation of haemophilic joint diseases because it enables the imaging of soft-tissue lesions in the joints and bone-cartilage lesions. We aimed to determine the correlation between US evaluations and clinical assessments performed using HJHS 2.1 and to evaluate their respective characteristics in assessing early haemophilic arthropathy. A total of 178 joints (32 knees, 85 elbows, and 61 ankles) in 45 haemophilia A patients (median age, 10 years; range, 6-15) were assessed using US and HJHS 2.1. Ultrasonographic scoring was performed in consensus assessments by one imager by using the US scores. The total HJHS 2.1 and US scores showed a strong correlation (rS=0.651, P=0.000, CI: 0.553-0.763), with an excellent correlation for the elbows (rS=0.867, P=0.000, CI: 0.709-0.941) and a substantial correlation for the knees (rS=0.681, P=0.000, CI: 0.527-0.797). The correlation for the ankles was relatively moderate (rS=0.518, P=0.000, CI: 0.308-0.705). Nine subjects (15.5%) without abnormalities, as indicated by HJHS 2.1, showed haemophilic arthropathy in US scoring. All nine joints showed moderate (1/9) to severe (8/9) synovial thickening in the ankle (5/9) and elbow joints (4/9). In contrast, 50 joints (50.5%) showed normal US scores and abnormal changes as indicated by HJHS 2.1. S scores correlated well with HJHS 2.1 for overall and individual joints. US could identify some early pathological changes in joints showing normal clinical findings, but still cannot replace the HJHS; however, it can serve as an imaging examination complementing HJHS 2.