This cross-sectional survey aims to determine whether cotinine exposure would enhance the relationship between blood manganese (Mn) and non-alcoholic fatty liver disease (NAFLD) in children using the NHANES database. Restricted cubic spline (RCS) and logistic regression analyses were adopted to determine the potential relationship. Besides, we tested the robustness of the results by performing trend tests and subgroup analyses. The study finally enrolled 866 children aged 18 years and below. Blood Mn was linearly linked to NAFLD and the risk of NAFLD was increased with the blood Mn elevation (P < 0.05). There was a notable relationship between blood Mn and NAFLD in crude model 1, which was still significant upon adjustment of all the identified covariates (all P < 0.05). Under Mn exposure, the cotinine-exposed group had a higher risk of NAFLD than the cotinine-unexposed group. In conclusion, blood Mn level is an independent risk factor for pediatric NAFLD, and cotinine exposure can enhance this relationship to some degree. Therefore, reducing cotinine exposure may alleviate detrimental consequences related to exposure to heavy metals in children.