In prior research, perceived low neighborhood social cohesion (nSC) has been associated with prevalence of type 2 diabetes mellitus (T2DM); however, few studies have investigated the nSC-T2DM relationship among a large, racially/ethnically diverse, and nationally representative sample of the U.S. population. We used National Health Interview Survey (2013–2018) data to determine overall, age-, sex/gender-, and racial/ethnic-specific associations between nSC and T2DM among 170,432 adults. Self-reported nSC was categorized as low, medium, and high. T2DM was determined by participants being told they had diabetes by a health professional. We used Poisson regression with robust variance to estimate prevalence ratios (PRs) and 95% confidence intervals (CI) while adjusting for confounders. Mean age was 47.4 ± 0.1 years, 52% were women, and 69% self-identified as Non-Hispanic (NH)-White. Low vs. high nSC was associated with a higher prevalence of T2DM (PR = 1.22 [95% CI: 1.16–1.27]), after adjustment. A higher prevalence of T2DM was observed among participants 31–49 years old who perceived low vs. high nSC (PR = 1.36 [95% CI: 1.20–1.54]) and among participants ≥50 years old (PR = 1.18 [95% CI: 1.13–1.24]). Hispanic/Latinx women 18–30 years old in neighborhoods with low vs. high social cohesion had a higher prevalence of T2DM (PR = 3.70 [95% CI: 1.40–9.80]), whereas NH-Black women 18–30 years old in neighborhoods with medium vs. high social cohesion had a lower prevalence of T2DM (PR = 0.35 [95% CI: 0.14–0.89]). Our findings support the literature by demonstrating an association between neighborhood environment and T2DM as well as extend it by identifying determinants for intervention for T2DM.