Objective The aim of this study was to: (a) screen a large group of unselected patients with juvenile systemic lupus erythematosus for anti-aquaporin 4 antibodies (AQP4-Ab); (b) identify clinical and laboratory predictors of the presence of AQP4-Ab positivity in juvenile systemic lupus erythematosus. Methods Sera from 90 patients with juvenile systemic lupus erythematosus were tested for the presence of AQP4-Ab using a cell-based assay. Demographics, clinical and immunological features, treatment received were summarized. Fisher’s exact test was used to identify clinical predictors of positivity for AQP4-Ab. Results Five of 90 (5.5%) patients tested positive for AQP4-Ab, all of which had neurological involvement, mainly transverse myelitis and optic neuritis. AQP4-Ab-positive patients were more likely to have neurological symptoms ( P = 0.002), less likely to experience dermatological manifestations ( P = 0.045), and less likely to have detectable anti-dsDNA antibodies ( P = 0.022). These patients were also more likely to have received anti-epileptic ( P = 0.023) and anti-coagulant ( P = 0.007) drugs. Conclusions The findings of this study indicate that some patients with juvenile systemic lupus erythematosus develop antibodies against aquaporin-4 and may be at risk of developing a neurological clinical phenotype. We suggest that all juvenile systemic lupus erythematosus patients should be systematically screened for the presence of AQP4-Ab and this may help identify a high risk for neurological involvement in juvenile systemic lupus erythematosus.