Background: Artificial Light at Night (ALAN) is an emerging health risk factor that has been linked to a wide range of adverse health effects. Recent study suggested that disadvantaged neighborhoods may be exposed to higher levels of ALAN. Understanding how social disadvantage correlates with ALAN levels is essential for identifying the vulnerable populations and for informing lighting policy. Methods: We used satellite data from the National Aeronautics and Space Administration’s (NASA) Black Marble data product to quantify annual ALAN levels (2012–2019), and the Center for Disease Control and Prevention’s (CDC) Social Vulnerability Index (SVI) to quantify social disadvantage, both at the US census tract level. We examined the relationship between the ALAN and SVI (overall and domain-specific) in over 70,000 tracts in the Contiguous U.S., and investigated the heterogeneities in this relationship by the rural-urban status and US re­ gions (i.e., Northeast, Midwest, South, West). Results: We found a significant positive relationship between SVI and ALAN levels. On average, the ALAN level in the top 20% most vulnerable communities was 2.46-fold higher than that in the 20% least vulnerable commu­ nities (beta coefficient (95% confidence interval) for log-transformed ALAN, 0.90 (0.88, 0.92)). Of the four SVI domains, minority and language status emerged as strong predictors of ALAN levels. Our stratified analysis showed considerable and complex heterogeneities across different rural-urban categories, with the association between greater vulnerability and higher ALAN primarily observed in urban cores and rural areas. We also found regional differences in the association between ALAN and both overall SVI and SVI domains. Conclusions: Our study suggested ALAN as an environmental justice issue that may carry important public health implications. Funding National Aeronautics and Space Administration.