Importance: The COVID-19 pandemic has negatively affected adult mental health (MH), with racial and ethnic minoritized groups disproportionately affected. Objective: To examine changes in adult MH-related emergency department (ED) visits into the Delta variant pandemic period and identify changes and inequities in these visits before and during COVID-19 case surges. Design, Setting, and Participants: This epidemiologic cross-sectional study used National Syndromic Surveillance Program data from US adults aged 18 to 64 years from 1970 to 2352 ED facilities from January 1, 2019, to August 14, 2021. All MH-related ED visits and visits related to 10 disorders (ie, anxiety, depressive, bipolar, schizophrenia spectrum, trauma- and stressor-related, attention-deficit/hyperactivity, disruptive behavioral and impulse, obsessive-compulsive, eating, and tic disorders) were identified. Exposures: The following periods of MH-related ED visits were compared: (1) high Delta variant circulation (July 18-August 14, 2021) with a pre-Delta period (April 18-May 15, 2021), (2) after a COVID-19 case peak (February 14-March 13, 2021) with during a peak (December 27, 2020-January 23, 2021), and (3) the Delta period and the period after a COVID-19 case peak with the respective corresponding weeks during the prepandemic period. Main Outcomes and Measures: ED visits for 10 mental disorders and all MH-related visits. Results: This cross-sectional study included 107 761 319 ED visits among adults aged 18 to 64 years (59 870 475 [56%] women) from January 1, 2019, to August 14, 2021. There was stability in most MH-related ED visit counts between the Delta and pre-Delta periods (percentage change, -1.4% to -7.5%), except for eating disorders (-11.9%) and tic disorders (-19.8%) and after a COVID-19 case peak compared with during a peak (0.6%-7.4%). Most MH-related ED visit counts declined in the Delta period relative to the prepandemic period (-6.4% to -30.7%); there were fluctuations by disorder when comparing after a COVID-19 case peak with the corresponding prepandemic period (-15.4% to 11.3%). Accounting for ED visit volume, MH-related ED visits were a smaller proportion of visits in the Delta period compared with the pre-Delta period (visit ratio, 0.86; 95% CI, 0.85-0.86) and prepandemic period (visit ratio, 0.80; 95% CI, 0.79-0.80). After a COVID-19 case peak, MH-related ED visits were a larger proportion of ED visits compared with during a peak (visit ratio, 1.04; 95% CI, 1.03-1.04) and the corresponding prepandemic period (visit ratio, 1.11; 95% CI, 1.11-1.12). Of the 2 510 744 ED visits included in the race and ethnicity analysis, 24 592 (1%) were American Indian or Alaska Native persons, 33 697 (1%) were Asian persons, 494 198 (20%) were Black persons, 389 740 (16%) were Hispanic persons, 5000 (0.2%) were Native Hawaiian or Other Pacific Islander persons, and 1 172 683 (47%) were White persons. There was between- and within-group variation in ED visits by race and ethnicity and increases in selected disorders after COVID-19 peaks for adults aged 18 to 24 years. Conclusions and Relevance: Results of this cross-sectional study suggest that EDs may have increases in MH-related visits after COVID-19 surges, specifically for young adults and individual racial and ethnic minoritized subpopulations. Public health practitioners should consider subpopulation-specific messaging and programmatic strategies that address differences in MH needs, particularly for those historically marginalized.