Aim To evaluate the quality of oral fluorescein angiography (FA) in relation to food intake. Methods This is an observational, case-crossover study. We collected information from patients undergoing routine oral FA for retinal disease at the Shiley Eye Institute. Eighty patients (160 eyes) were analysed. Fasting and non-fasting images of the same patient were recorded, compared and analysed for different image quality parameters and clinical relevance by experienced retina specialists. Results When analysing the images, intergrader agreement was moderate to good with a Kappa averaging 0.60 (0.5–0.85). When patients were fasting pre-imaging, better angiography quality scores were achieved when compared with images taken when patients were non-fasting (mean 0.84 vs 0.72, p<0.001). Multivariate analysis showed that non-fasting patients with higher body mass index had the worst scores. Other clinical parameters, such as staining of drusen, staining of disciform scars or central and peripapillary atrophy, were also significantly better during the pre-fasting exam (p<0.001). Oral FA was approximately 22% faster (time to fluorescein dye appearance) under fasting conditions than non-fasting (mean±SD, minutes, 18.7±6.9 vs 25.14±8.1, p<0.001). Conclusion Fasting oral FA provided significantly better quality images as well as faster optimal imaging times when compared with non-fasting oral FA. By improving its overall quality, oral FA could be a useful adjunctive examination to optical coherence tomography (OCT) and OCT angiography in patients who require FA studies but who have difficult access or refuse an invasive procedure.