BACKGROUND For patients with idiopathic or multiple sclerosis (MS)-associated optic neuritis (ON), the largest multicenter clinical trial (Optic Neuritis Treatment Trial [ONTT]) showed excellent visual outcomes and baseline high-contrast visual acuity (HCVA) was the only predictor of HCVA at 1 year. We aimed to evaluate predictors of long-term HCVA in a modern, real-world population of patients with ON and compare with previously published ONTT models. METHODS We performed a retrospective, longitudinal, observational study at the University of Michigan and the University of Calgary evaluating 135 episodes of idiopathic or MS-associated ON in 118 patients diagnosed by a neuro-ophthalmologist within 30 days of onset (January 2011-June 2021). Primary outcome measured was HCVA (Snellen equivalents) at 6-18 months. Multiple linear regression models of 107 episodes from 93 patients assessed the association between HCVA at 6-18 months and age, sex, race, pain, optic disc swelling, symptoms (days), viral illness prodrome, MS status, high-dose glucocorticoid treatment, and baseline HCVA. RESULTS Of the 135 acute episodes (109 Michigan and 26 Calgary), median age at presentation was 39 years (interquartile range [IQR], 31-49 years), 91 (67.4%) were women, 112 (83.0%) were non-Hispanic Caucasians, 101 (75.9%) had pain, 33 (24.4%) had disc edema, 8 (5.9%) had a viral prodrome, 66 (48.9%) had MS, and 62 (46.6%) were treated with glucocorticoids. The median (IQR) time between symptom onset and diagnosis was 6 days (range, 4-11 days). The median (IQR) HCVA at baseline and at 6-18 months were 20/50 (20/22, 20/200) and 20/20 (20/20, 20/27), respectively; 62 (45.9%) had better than 20/40 at baseline and 117 (86.7%) had better than 20/40 at 6-18 months. In linear regression models (n = 107 episodes in 93 patients with baseline HCVA better than CF), only baseline HCVA (β = 0.076; P = 0.027) was associated with long-term HCVA. Regression coefficients were similar and within the 95% confidence interval of coefficients from published ONTT models. CONCLUSIONS In a modern cohort of patients with idiopathic or MS-associated ON with baseline HCVA better than CF, long-term outcomes were good, and the only predictor was baseline HCVA. These findings were similar to prior analyses of ONTT data, and as a result, these are validated for use in conveying prognostic information about long-term HCVA outcomes.