Background Physical activity is associated with positive health biomarker profiles during pregnancy and resultant birth outcomes. Furthermore, there is some evidence to suggest that physical activity both prior to and during pregnancy may reduce the risk of gestational diabetes. In this study we aim to investigate the relationship between physical activity and biomarkers associated with gestational diabetes (GDM) risk and clinical diagnoses of GDM. Methods Participants were pregnant women recruited to the Born in Bradford cohort study who completed phase 1 or 2 of the baseline questionnaire and provided a fasting blood sample between approx. 26–28 weeks of gestation, and had a singleton birth. Where mothers had two births during the study period, only the first was included in the present analyses. Physical activity level was measured using the General Practice Physical Activity Questionnaire (GPPAQ) and questions about usual walking speed. The relationship of self-reported physical activity, using both of the two definitions above, with levels of fasting blood glucose was evaluated using multivariate linear regression. GDM risk was assessed using logistic regression analyses. Models were adjusted for other key covariables including age, ethnicity, body mass index and parity. Results Data were available for 6119 maternal participants. White British women reported higher levels of physical activity and physical function (i.e. higher self-reported average walking speed) relative to their Pakistani and Other ethnic group peers. Despite the higher GPPAQ scores in the White British group, over 67 per cent still fell into the inactive or moderately inactive category. Higher levels of activity as measured by GPPAQ scores was not associated with fasting blood glucose or odds of gestational diabetes. Walking speeds were associated with lower levels of fasting glucose, and remained so after adjustment for other relevant covariables (−0.04 (−0.08, –0.05). Adjusted odds of gestational diabetes diagnosis were also lower in those reporting the highest self-reported walking speeds OR 0.49 (0.27–0.87). Conclusion Faster self-reported walking speeds, which may result from greater residual fitness prior to pregnancy, were shown to be associated with lower fasting blood glucose levels and lower odds of a gestational diabetes diagnosis. These results provide first time evidence that self-reported walking speed, an indicator of functional reserve, is associated with positive pregnancy related biomarker profiles and require independent confirmation. Given the lower physical activity and functional profile of Pakistani women, there remains high potential for behaviour change interventions in this population group. To realise this ambition further research must focus on understanding specific cultural and socio-economic barriers to implementation.