This study examined how race/ethnicity, sex/gender, and sexual orientation intersect under interlocking systems of oppression to socially pattern depression among US adults. With cross-sectional data from the 2015-2020 National Survey on Drug Use and Health (n = 234 722), we conducted a design-weighted, multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) under an intersectional framework to predict past-year and lifetime major depressive episodes (MDEs). With 42 intersectional groups constructed from 7 race/ethnicity, 2 sex/gender, and 3 sexual orientation categories, we estimated age-standardized prevalence and excess or reduced prevalence attributable to 2-way or higher interaction effects. Models revealed heterogeneity across groups, with prevalence ranging from 1.9% to 19.7% (past-year) and 4.5% to 36.5% (lifetime). Approximately 12.7% (past year) and 12.5% (lifetime) of total individual variance was attributable to between-group differences, indicating key relevance of intersectional groups in describing the population distribution of depression. Main effects indicated, on average, that people who were White, women, gay/lesbian, or bisexual had greater odds of MDE. Main effects explained most between-group variance. Interaction effects (past year: 10.1%; lifetime: 16.5%) indicated another source of heterogeneity around main effects average values, with some groups experiencing excess or reduced prevalence compared with main effects expectations. We extend the MAIHDA framework to calculate nationally representative estimates from complex sample survey data using design-weighted, Bayesian methods. This article is part of a Special Collection on Mental Health.