In children with juvenile idiopathic arthritis (JIA), the temporomandibular joint (TMJ) can be involved. To prevent TMJ damage due to inflammation, early recognition is important, for which contrast-enhanced magnetic resonance imaging (MRI) is the gold standard. In this study, the interobserver reliability and construct validity of the Juvenile Idiopathic Arthritis Magnetic Resonance Scoring System for Temporomandibular Joints (JAMRIS-TMJ) was assessed. Two radiologists independently examined 38 MRIs using the JAMRIS-TMJ scoring system. Inter-observer reliability was assessed by Cohen’s (weighted) kappa (κ), 95% confidence intervals (CIs) and absolute agreement (%). Construct validity was assessed by correlation between the JAMRIS-TMJ items and TMJ involvement, active maximum interincisal mouth opening (AMIO), and anterior maximum voluntary bite force (AMVBF). The interobserver reliability for the JAMRIS-TMJ items varied from poor to good (κ=0.18–0.61). Joint enhancement had the highest reliability (κ=0.61). Correlations were found between TMJ involvement, AMIO, and the JAMRIS-TMJ items, although variation between radiologists and TMJ side existed. No correlation was found between AMVBF and the JAMRIS-TMJ items for both radiologists. The strongest correlations were found between most of the JAMRIS-TMJ items and AMIO. Our findings support the utility of AMIO as a clinical measure of TMJ status in children with JIA.