Objective: To study the association between mode of conception and risk of preterm birth (PTB), including, spontaneous and provide-initiated PTB. Design: Retrospective cohort study Setting and Population: Population-based cohort included all singleton livebirths and stillbirth in Ontario, Canada, 2006-2014. Methods: Mode of conception comprised (i) subfertility without infertility treatment (N = 68,822); (ii) non-invasive infertility treatment (ovulation induction +/- intrauterine insemination) (N = 9024); or (iii) Invasive infertility treatment (in vitro fertilization [N = 8038) – each compared to births by unassisted conception (N = 646,926). Modified Poisson regression generated risk ratios (RR). Confounding was handled by inverse probability of treatment weighting using a propensity that included maternal demographics and pre-existing conditions. Main Outcome Measure: PTB < 37 completed weeks’ gestation. PTB was further categorized as spontaneous PTB or provider-initiated PTB. Results: PTB occurred among 6.0% of births by unassisted conception, 7.7% with subfertility, 8.0% with non-invasive infertility treatment, and 10.8% following invasive infertility treatment. The RR of provider-initiated PTB was higher in women with subfertility (RR 1.23, 95% CI 1.16-1.31), non-invasive infertility treatment (RR 1.48, 1.29-1.69) and invasive infertility treatment (RR 2.35, 2.09-2.64) – each relative to births by unassisted conception. The corresponding RR for spontaneous PTB were 1.15 (95% CI 1.10-1.19), 1.19 (95% CI 1.09-1.31) and 1.40 (95% CI 1.27-1.53). Conclusions: Subfertility, and receipt of infertility treatment, are each associated with a higher risk of PTB, especially provider-initiated PTB. Strategies are needed to reduce the underlying indications to deliver these women before term.