Thyroid hormones (TH) are critical in early neurodevelopment, but few studies have examined whether neonatal TH levels influence risk of autism spectrum disorder (ASD). This study linked California neonatal screening data with live birth and Department of Developmental Services records to examine newborn TH levels in relation to ASD. Thyroxine (T4) and thyroid‐stimulating hormone (TSH) levels were measured in newborn bloodspots as part of routine screening, in 1996 and 2002, respectively. Mean levels of T4 and TSH were compared between ASD cases and non‐cases. Four hundred forty‐seven thousand, fifty‐nine screened, singleton births from 1996 and 446,424 from 2002 were examined, including 4,818 ASD cases. Binomial regression, using categories of T4 and TSH percentiles was used to calculate crude and adjusted risk ratios (RR). Covariates included maternal and child factors, gestational age, and age at blood draw. No significant associations were found with TSH levels and ASD in crude or adjusted analyses. ASD cases had significantly lower mean T4 levels than non‐cases, but this association was no longer significant in adjusted analyses (RR in individuals in lowest 5th percentile of T4 levels = 1.13, 95% 0.93–1.37). However, this association appeared stronger in certain subgroup analyses, particularly among neonates with blood draw ≥48 hr from birth (RR = 1.67, 95% CI 1.08, 2.60), when TH levels become more stable. Thus, results from this large, population‐based study did not suggest strong associations between neonatal TH and ASD, but certain subgroups of newborns with the lowest T4 levels may have modestly increased ASD risk. Autism Res 2016. © 2016 International Society for Autism Research, Wiley Periodicals, Inc. Autism Res 2017, 10: 585–592. © 2016 International Society for Autism Research, Wiley Periodicals, Inc.