Background <p>The health benefits of exercise have been well-established, but the specific role of yoga in multimorbidity was less understood. This study aimed to examine the association between yoga and multimorbidity in comparison with other exercises or non-exercise.</p> Methods <p>This cross-sectional study pooled nationally representative surveys from Behavioral Risk Factor Surveillance System (BRFSS) (2015, 2017, and 2019), and classified 279,885 participants aged 45 + years as yoga practitioners, other exercisers, and non-exercisers. Multinominal and binary logistic regressions were separately used for association of yoga with multimorbidity (coexistence of ≥ 2 conditions) and individual chronic conditions. Potential effect modification by age, sex, education level, and race/ethnicity was examined. We further assessed the dose-response association of duration and frequency of yoga practice with multimorbidity. </p> Results <p>The adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for yoga practitioners (<i>n</i> = 5081) vs. other exercisers (<i>n</i> = 192,718) were 0.69 (95% CI 0.55–0.86) for coexistence of two conditions, and 0.72 (95%CI 0.58–0.89) for coexistence of ≥ 3 conditions. The associations were stronger (ORs ranged from 0.43 to 0.52; <i>P</i> &lt; 0.0001 for all) when comparing yoga practitioners with non-exercisers (<i>n</i> = 82,086), which were more pronounced in women relative to men (<i>P</i>-interaction &lt; 0.05). Increased duration or frequency of yoga practice was associated with lower odds of multimorbidity in a dose-response manner (<i>P</i> for trend&lt;0.0001 for all). Similar inverse associations were observed between yoga and individual chronic conditions.</p> Conclusion <p>Yoga was associated with lower odds of multimorbidity in middle-aged and older adults, relative to other exercises or non-exercise.</p>