Arsenic occurs as a natural contaminant of drinking water supplies in arsenic endemic areas, posing a threat to public health. Our aim was to evaluate the relationship between urinary arsenic concentrations and spontaneous pregnancy loss in a population with low-moderate level drinking water arsenic exposure (mostly <10 μg/L). We enrolled 150 women with incident spontaneous pregnancy losses and 150 controls with ongoing pregnancies matched by gestational age. We measured arsenic species in urine using high performance liquid chromatography paired to inductively coupled plasma mass spectrometry. Urinary arsenic was not related to spontaneous pregnancy loss in conditional logistic regression models adjusted for confounding factors. However, a 10 μg/L increase in urinary arsenic (III + V) salt concentrations was associated with 8.00-fold (95% CI: 0.68, 3.35 × 105) greater odds of spontaneous loss among women using prenatal vitamins in an interaction model (P for interaction = 0.07), although the effect estimate was imprecise. In an additional interaction model, prenatal vitamin use was associated with lower odds of loss (OR = 0.30; 95% CI: 0.13, 0.66), although the association was diminished in the presence of a 10 μg/L increase in urinary inorganic arsenic (OR = 0.44; 95% CI: 0.70, 3.22). Total urinary arsenic was associated with 1.48-fold (95% CI: 0.20, 11.35) greater odds for loss among women with urinary cotinine >50 μg/L in another interaction model (P for interaction = 0.07). These results suggest a potential modest increase in the odds of pregnancy loss associated with increased total urinary arsenic among women smoking during pregnancy (urinary cotinine >50 μg/L). Prenatal vitamin use may act as a protective factor for arsenic exposure associated pregnancy loss, but appears to be less protective with increasing urinary inorganic arsenic concentrations.