OBJECTIVE: Prematurity affects children's functioning in domains such as cognitive skills and math skills. However, there is limited research examining math skills as gestational age (GA) increases weekly. We aimed at determining, in a population-based cohort of children aged 8 to 9 years, the association between the GA spectrum and math skills and working memory's (WM's) role in this relationship. METHODS: Children aged 8 to 9 years in 2014 (n = 1761), from 44 primary schools in a metropolitan city, were recruited through the Memory Maestros study in 2012. GA was measured using the parent report. Math skills were measured using the Wide Range Achievement Test 4 math computation subtest (mean, 100; SD, 15), and WM was measured using 2 subtests of the Automated Working Memory Assessment. The method of analysis was multivariate linear regression, with adjustment for both WM measures and social risk indicators as potential confounders. RESULTS: A total of 1168 children (66%) had analyzable data. Compared with children of ≤34-week gestation, there was a mean difference increase of 0.48 in math standard scores per weekly increase in GA (95% confidence interval, 0.11-0.85%; p = 0.01). This equates to a difference of 6.29 (0.42 SD) standard math scores between the ≤34- and ≥42-week group. The relationship was independent of WM. CONCLUSION: Weekly increases in GA beyond 34 weeks are associated with higher mathematics achievement scores for children aged 8 to 9 years, independent of the child's WM. This assists with identification of individuals at risk of poorer math skills.