Background: Efferent visual dysfunction in children could lead to impaired quality of life at home and school. Eye-tracking can detect subtle efferent dysfunction missed on bedside examination but has not been validated in the pediatric multiple sclerosis population. Objective: We sought to determine the feasibility of eye-tracking in children and associations with multiple sclerosis. Methods: Participants meeting criteria for pediatric multiple sclerosis without acute efferent vision abnormalities and healthy controls were recruited. Multiple sclerosis participants underwent a clinical assessment and saccade and antisaccade testing paradigms. Intraclass correlation coefficients were generated for intertest repeatability. Adjusting for age and intereye correlations, generalized estimating equations compared latencies with case status, Expanded Disability Status Scale and Symbol Digit Modalities Test (SDMT) scores. Results: We eye-tracked 15 children with multiple sclerosis (n = 30 eyes, mean age 15.6 ± 2.1, mean disease duration 3.9 years, median Expanded Disability Status Scale 1.5) compared to 6 healthy controls (n = 12 eyes, age 14.3 ± .95). The intraclass correlation coefficient for repeated trials was 0.85. Adjusting for age, saccadic latency was 60 milliseconds (ms) longer for cases than controls (95% confidence interval = 26.4, 93.8; P = .0005). For antisaccadic latency, we observed a similar trend of 60 ms longer for cases than controls ( P = .06). Conclusion: Eye-tracking is a short noninvasive examination, and high intertest repeatability supports use of eye-tracking technology in pediatric multiple sclerosis. Longer saccadic latencies were seen in children with multiple sclerosis despite short disease duration and low Expanded Disability Status Scale scores.