A cognitive theory of visual expertise is proposed to model the visual information processing of domain experts. During theory development, we integrated empirical evidence from multiple methods of inquiry—including qualitative observations, qualitative interviews, eye tracking and neuroimaging—used in research on learning and instruction in the professions. Devoted to understand the skilled perception, interpretation and evaluation of visual information in work tasks, the theory is based on three underlying assumptions: extended capacity, knowledge-based processing and practice-based interaction. Expert professional vision is described as a set of eight processes in the visual register and long-term working memory; these processes are selecting visual information, ignoring visual information, knowledge-based noticing, extending the visual field through parafoveal processing, organising image chunks, integrating, using visual practices to interact with the environment and monitoring. We discuss the educational and methodological implications of the theory as they relate (a) to the expertise development of novices through training and deliberate practice in vision-intensive workplaces—including, but not limited to, the arts, education, sports, medicine, transportation and programming—and (b) to the use of mixed methods when studying professional vision.