Background: Migraine is a common and disabling neurological disorder with complex symptoms, which may include eating-related symptoms before, during, and after headache (HA) pain. Objectives: This study aimed to use a novel case-crossover design to assess dietary intake in temporal relation to migraine attacks. Methods: Data were collected from 25 people with migraine over a 28 d period, consisting of twice weekly dietary recalls combined with daily HA diaries assessing migraine symptoms. Micro- and macronutrient intakes were assessed, along with dietary quality. Generalized estimating equations assessed relationships between migraine attack phases and day-level dietary intake. Results: Participants included 23 females and 2 males, with a mean (M) age of 31 ± 12 y and with a mean migraine disability assessment score in the severe range at 28.6 ± 20.2. Vitamin D and choline had the lowest frequencies of participants meeting the Dietary Reference Intakes. During the prodrome phase, the Healthy Eating Index (HEI) component score for total protein foods [M: 4.6; 95% confidence interval (CI): 4.4, 4.9], calories from total protein (M: 19.6%; 95% CI: 17.1%, 22.0%), and animal protein (M: 13.4%; 95% CI: 10.8%, 16.1%) were significantly higher than on interictal days [(M: 4.1; 95% CI: 3.9, 4.4), (M: 15.7%; 95% CI: 14.5%, 16.9%), and (M: 9.6%; 95% CI: 8.4%, 10.8%), respectively]. On HA days with mild pain, total HEI scores [(M: 61.9; 95% CI: 53.8, 69.9), total vegetable HEI component (M: 4.4; 95% CI: 3.9, 5.0)], greens/beans component (M: 3.8; 95% CI: 2.7, 4.9) and fiber density (M: 15.6 g/1000 kcal; 95% CI: 12.8, 18.3 g/1000 kcal)], were higher than on HA days with severe pain [(M: 52.6; 95% CI: 47.5, 57.8), (M: 3.1; 95% CI: 2.5, 3.8), (M: 2.3; 95% CI: 1.4, 3.1), and (M: 11.1; 95% CI: 9.2, 13.0), respectively]. Conclusions: A study design utilizing a novel combination of dietary recalls and HA diaries enables the observation of day-level differences in protein intake and dietary quality in relation to phase of the migraine attack and HA pain severity. The methodology is successful in consistently gathering detailed data that supports future designs, which represent diverse populations with migraine.