To explore the association of marijuana use with mean plasma fasting insulin levels and homeostasis model assessment of insulin resistance (HOMA-IR) score in obese adults with different HOMA-IR.The National Health and Nutrition Examination Survey (NHANES) is a survey research program designed to assess the health and nutrition status of individuals in the United States and to track changes over time. We abstracted data from NHANES 2009-2016. We estimated the minimal lifetime marijuana use (MLU) using the duration of regular exposure and the frequency of use. We assessed the association of MLU and both plasma fasting insulin and HOMA-IR score using log-linear regression models.A total of 65,209 obese individuals aged 18 to 59 years were included. In obese individuals who never used marijuana (reference), the mean value (± standard deviation) was 19.0 (± 12.8) μU/mL for plasma fasting insulin and 4.78 (± 3.49) for HOMA-IR. In individuals with HOMA-IR < 2.13 or ≥ 5.72, we found no association of marijuana use with HOMA-IR. In those with HOMA-IR < 5.72, the highest tertile of MLU (i.e., ≥ 1799 times) was associated with 12% decrease (95% confidence intervals, 4-19%) in the fasting insulin and 10% decrease in HOMA-IR (95% CI 1-19%), as compared with their counterparts who never used marijuana. In those with HOMA-IR ≥ 2.13, we found a marked impact of marijuana use only in adults who used marijuana ≥ 1799 times, with 13% decrease (95% CI 5-19%) in fasting insulin and 10% decrease (95% CI 3-18%) in HOMA-IR score.Marijuana use is associated with reduced fasting insulin levels and HOMA-IR score in US obese adults with HOMA-IR ≥ 2.13, but not in those with HOMA-IR < 2.13 or ≥ 5.72. The impact of marijuana use is the greatest after long-term exposure and is independent of BMI.