Background: Early-life exposure to nonessential (toxic) and essential trace elements can influence child development. Although infant formula powders and the water used to reconstitute them can contain higher concentrations of many elements compared with human milk, the influence of feeding mode on reliable biomarkers of infant exposure has rarely been demonstrated. Methods: We evaluated associations between urinary biomarkers and feeding mode (exclusively human milk, exclusively formula, or combination-fed) for four toxic (arsenic, cadmium, nickel, and uranium) and three essential elements (cobalt, molybdenum, and selenium) using general linear models. Results: A total of 462 participants from the rural New Hampshire Birth Cohort Study were on average 6 weeks old between July 2012 and March 2019 and had urine samples, 3-day food diaries, and relevant covariate data available. In adjusted models, urinary arsenic was 5.15 (95% confidence interval = 4.04, 6.58), molybdenum was 19.02 (14.13–25.59), and selenium was 1.51 (1.35–1.68) times higher in infants fed exclusively with formula compared with infants fed exclusively with human milk. By contrast, urinary uranium was 0.59 (0.46–0.75) and cobalt was 0.78 (0.65–0.95) times lower with formula feeding than human milk feeding. Conclusion: Our findings suggest that infant exposure to several potentially toxic elements varies by feeding mode, as concentrations of reliable urinary biomarkers were higher with formula or human milk, depending on the element. Importantly, exposure to arsenic increased with household tap water arsenic regardless of feeding mode, suggesting that all infants could be at risk in populations with high concentrations of arsenic in drinking water.