Education is widely associated with better health outcomes, but whether this relationship is causal remains unclear, especially in low-income countries. This study examines the causal effect of maternal education on child nutrition and dietary diversity in Kenya. Using a fuzzy regression discontinuity design, with exposure to the 2003 school reform as an instrument for educational attainment, the analysis finds that an additional year of maternal schooling significantly improves children’s nutritional status and dietary diversity. Positive effects are observed for both boys and girls, although they are slightly stronger for boys. The results are robust to a range of sensitivity tests. Further analysis suggests that maternal education improves child health through several channels, including delayed first birth, fewer children under age five, higher paternal education, greater use of prenatal care, increased access to information through newspapers and television, improved literacy, and higher employment rates among mothers. These findings indicate that expanding educational opportunities for girls can be an effective policy tool for improving child nutrition and feeding practices in developing countries such as Kenya.