We conducted a population-based case-control study, using the 1984–1989 linked Washington State birth and death certificate data, to determine whether maternal prior spontaneous pregnancy loss was associated with the risk of sudden infant death syndrome (SIDS). We identified 87Z autopsied SIDS cases and 3,704 controls matched to cases on year of birth. To reduce the confounding from unmeasured socio-economic and behavioral factors, we analyzed the data separately for infants with and without father's race provided on the birth certificate. Among infants whose father's race was known, those of multiparous women with a prior spontaneous pregnancy loss before 20 weeks gestational age were at reduced risk of SIDS, after adjustment for maternal age, prenatal smoking, and gravidity. An increased number of prior pregnancy losses was associated with a further reduction in the risk of SIDS: one prior pregnancy loss was associated with a 25% decrease [odds ratio (OR) = 0.75; 95% confidence interval (CI) = 0.52–1.08], and two or more prior pregnancy losses was associated with a 47% decrease (OR = 0.53; 95% CI = 0.30–0.94). There was no such association, however, among infants of women with no prior livebirth. Similar associations were observed between prior spontaneous pregnancy loss during the entire gestational period and the risk of SIDS in offspring. To explain the different associations between maternal prior pregnancy loss and the risk of SIDS observed among infants of women with and without a prior livebirth, we hypothesize an upward bias resulting from the presence of a larger proportion of infants of human leukocyte antigen-compatible couples among first liveborn infants.