While an association between low socio-economic status (SES) and increased risk of stillbirth has been observed consistently over several decades, the pathways through which SES exerts these effects have not been established. Given that some key health-risk behaviours for stillbirth, including smoking and pre-pregnancy obesity, have strong relationships with SES, health-risk behaviours may serve as a channel through which low SES contributes to stillbirth outcomes. The objective of this study was to estimate the proportion of the relationship between low SES and the occurrence of stillbirth that is explained by health-risk behaviours in populations of Eastern Ontario and Nova Scotia (112 stillbirth cases and 398 controls). Both area and individual level influences of SES were assessed. The study population consisted of 112 cases (women delivering stillborn infants) and 398 controls. Odds ratios and 95% confidence intervals estimated by multivariable logistic regression were used to approximate relative risks. The contribution of health-risk behaviours to relationships between SES and stillbirth was assessed by a change in the relative risk estimate following omission of each health-risk behaviour from the model. Of the three measures of individual level SES examined (household income, education, Blishen occupational index), only household income was a statistically significant predictor of stillbirth. After controlling for individual level SES, no community level SES effects were observed for stillbirth. Adjustments for key health-risk behaviours (smoking) resulted in an 18.5% reduction in the odds ratio estimate for low SES, from 3.31 to 2.79. This large unexplained SES effect that remained highlights the need for research into other potential pathways that may account for increased risk of stillbirth among those of lower SES.