Background: This study assessed the importance of mode of administration relative to other treatment attributes when selecting a preventive migraine therapy. Methods: Cross-sectional study among Canadian adults diagnosed with migraine with ≥5 monthly migraine days and tried ≥2 prescription migraine treatments (any kind/duration). Preferences for treatments varying in the following attributes were evaluated via a discrete choice experiment: speed of efficacy (effective in 24hr/1wk/3mo), duration of efficacy (wears off never/1wk/2 wks before next dose), mode of administration (infusion/auto-injection/cranial injections), administration setting (clinic/home), and administration frequency (1mo/3mo). Attribute-level preference weights were estimated using Hierarchical Bayes modeling. Results: Of 200 respondents, 142 experienced episodic migraine and 58 experienced chronic migraine. Preference weights confirmed that respondents’ most preferred treatments were those that provided fast and long-lasting efficacy (effective in 24hr = 0.59; wears off never = 1.07) and were offered via infusion (0.58) or auto-injection (0.47) over intracranial injection (-1.04). Respondents reported being moderately willing to receive infusions in either a home or clinic setting (1-6 Likert scale from “not at all” to extremely” willing). Conclusions: Second to speed and duration of efficacy, respondents were most concerned with mode of administration when selecting their preferred migraine preventive, suggesting that physicians should consider patient preferences in treatment decision-making.