The objective of this study was to determine the prevalence and predictors of yoga practice. This cross-sectional study was conducted as a part of a larger study that included yoga as a form of physical activity. Data were collected during April and August 2019 from the adult urban population of Bhubaneswar, India. This study was conducted using a cluster random sampling method. A representative sample (n = 1,203) of adults aged 18–59 years, irrespective of gender, was interviewed using a questionnaire adapted from the 2012 U.S. National Health Interview Survey, with the Epicollect5 handheld data-collection tool. Predictors of yoga practice were explored using multivariable logistic regression. The mean age of the participants was 35.19 ± 10.67 years, with 55.3% males. The majority were Hindu (93.62%) and belonged to the unre s e rved category (65.60%), people generally of higher relative socioeconomic status. The lifetime prevalence of yoga was 16.9%. Prevalence of any form of yoga (yoga, pranayama, or meditation), all forms of yoga (yoga, pranayama, and meditation), pranayama, and meditation was 17.0%, 10.7%, 14.3%, and 11.4%, respectively. After adjusting for confounders, female gender, Hindu religion, minimum of higher-secondary or graduate-level education, and having received advice from professionals for yoga practice had significantly higher odds of practicing yoga, and those of higher socioeconomic status had significantly lower odds of practicing yoga. We found a low prevalence of yoga. Sociodemographic characteristics like gender, religion, education, socioeconomic status, and other factors like learning yoga from professionals may be important predictors of continued yoga practice.