Objective This study aimed to determine the correlation between interpregnancy interval (IPI) and neurodevelopmental delay in infants and to explore the potential mediating role of maternal-fetal glucose metabolism. Design Population based cohort study. Setting China Participants A total of 2559 mother-infant pairs between 2018 and 2022. Methods The prospective birth cohort study included 2559 mother-infant pairs. The IPI is calculated by subtracting the gestational age of the current pregnancy from the interval at the end of the previous pregnancy. Maternal fasting venous blood collection at 24-28 weeks and cord blood collection at delivery. The association between IPI and neurodevelopment was determined by logistic regression. Main outcome measures Neurodevelopmental outcomes at 12 months in offspring were assessed by the Ages and Stages Questionnaire Edition 3 (ASQ-3). Results In our cohort, 14.0% had an IPI <12 months. IPI <12 months increased the failure of the communication domain, fine motor domain, and personal social domain of the ASQ [relative risks (RRs) with 95% confidence interval (CI): 1.75(1.13, 2.72); 1.77(1.13, 2.77); 1.51(1.01, 2.28)]. Maternal homeostasis model assessment of insulin resistance (HOMA-IR) and cord blood C-peptide was significantly associated with failure in the communication domain [RRs with 95% CI: 1.15(1.02, 1.31); 2.15(1.26, 3.67)]. The proportion of the association between IPI and failure of the communication domain risk mediated by maternal HOMA-IR and cord blood C-peptide was 14.4%. Conclusions IPI <12 months was associated with failing the communication domain in offspring. Maternal-fetal glucose metabolism abnormality may partially explain the neurodevelopmental delay caused by short IPI.