Background: The association between physical activity (PA), sedentary behavior, and incident diabetes has been assessed in whites but is less well investigated in multiethnic populations. Objective: To assess the association between PA, sedentary behavior, and incident diabetes in the Multi- Ethnic Study of Atherosclerosis. Research design and methods: Incident diabetes was assessed among adults without prevalent baseline diabetes (2000–2002) at 5 in-person examinations between 2002 and 2012. Baseline PA (moderate, vigorous, and exercise-specific; metabolic equivalents of task-hours/week) and sedentary behaviors (television watching, reading; hours/day) were assessed by questionnaire. HRs were estimated using Cox proportional hazard models. Results: Among 5829 adults (mean age 61.8 years, 54% female, 42% white, 12% Chinese-American, 26% African-American, 21% Hispanic-American), there were 655 incident diabetes cases (median follow-up 11.1 years). After adjustment, diabetes risk was lower in those with brisk or striding compared with none or casual walking pace (HR 0.67; 95% CI 0.54 to 0.84), higher levels of exercise PA (HR for highest vs lowest quartile 0.79; 95% CI 0.63 to 0.98), and any compared with no vigorous PA (HR 0.79; 95% CI 0.66 to 0.95). Race/ethnicity influenced the association of walking pace, exercise PA, and any vigorous PA on diabetes risk, which was only significant among whites. Total leisure sedentary behaviors (HR for highest vs lowest quartile 1.65; 95% CI 1.26 to 2.14) and television watching (HR for highest vs lowest quartile 2.68; 95% CI 1.38 to 5.21) were significantly associated with diabetes risk in multiethnic analyses and were influenced by race/ethnicity. Conclusions: These results confirm the importance of PA and sedentary behavior on diabetes risk in a multiethnic population and demonstrate potential variations across race/ethnic groups.