Preclinical evidence indicates the potential anti-tumor capabilities of cannabinoids in prostate cancer (PC). We undertook a cross-sectional study using National Survey on Drug Use and Health data from 2002 to 2020, involving 2503 participants in the USA. The independent variable was marijuana use status (current, former, never), while the dependent variable was self-reported PC (yes, no). Eleven other demographic variables were assessed as covariates. PC prevalence was lower among current marijuana users (46/145, 31.7%) and former users (323/1021, 31.6%) compared to non-users (534/1337, 39.9%, p < 0.001). PC prevalence was lower among users versus non-users in the elderly (≥65) (36.4% vs. 42.4%, p = 0.016) and non-Hispanic white subgroups (28.9% vs. 38.3%, p < 0.001). There were no significant PC prevalence differences between users and non-users in the younger population (50–64) or other race/ethnicity. In the multivariable analyses, former marijuana use was associated with lower PC compared to never using (odd ratio = 0.74, 95% CI 0.62–0.90, p = 0.001). Current use was also suggestive of reduced prevalence but was not statistically significant (odd ratio = 0.77, 95% CI 0.52–1.14, p = 0.198), possibly due to low sample size. Our findings from a large national survey provide additional data to link marijuana use with lower PC prevalence.