The evidence regarding a potential link of low-to-moderate iodine deficiency, selenium status, and cadmium exposure during pregnancy with neurodevelopment is either contradicting or limited. We aimed to assess the prenatal impact of cadmium, selenium, and iodine on children’s neurodevelopment at 4 years of age. The study included 575 mother–child pairs from the prospective “Rhea” cohort on Crete, Greece. Exposure to cadmium, selenium and iodine was assessed by concentrations in the mother’s urine during pregnancy (median 13 weeks), measured by ICPMS. The McCarthy Scales of Children’s Abilities was used to assess children’s general cognitive score and seven different sub-scales. In multivariable-adjusted regression analysis, elevated urinary cadmium concentrations (≥0.8 µg/L) were inversely associated with children’s general cognitive score [mean change: −6.1 points (95 % CI −12; −0.33) per doubling of urinary cadmium; corresponding to ~0.4 SD]. Stratifying by smoking status (p for interaction 0.014), the association was restricted to smokers. Urinary selenium was positively associated with children’s general cognitive score [mean change: 2.2 points (95 % CI −0.38; 4.8) per doubling of urinary selenium; ~0.1 SD], although the association was not statistically significant. Urinary iodine (median 172 µg/L) was not associated with children’s general cognitive score. In conclusion, elevated cadmium exposure in pregnancy of smoking women was inversely associated with the children’s cognitive function at pre-school age. The results indicate that cadmium may adversely affect neurodevelopment at doses commonly found in smokers, or that there is an interaction with other toxicants in tobacco smoke. Additionally, possible residual confounding cannot be ruled out.Electronic supplementary materialThe online version of this article (doi:10.1007/s10654-016-0151-9) contains supplementary material, which is available to authorized users.