A gaze-driven methodology for discomfort glare was developed and applied for glare evaluation. A series of user assessments were performed in an office-like test laboratory under various lighting conditions. The participants’ gaze responses were recorded by means of mobile eye tracking while monitoring photometric quantities relevant to visual comfort using HDR luminance imaging. The integration of the luminance images coupled with eye-tracking methods enabled us to use gaze-centred luminance distribution to have an accurate estimate of the light received at the eye. Using a novel gaze-driven approach, a unique database was created as a basis to investigate the gaze direction dependencies of visual comfort. Here we compare the proposed gaze-driven approach with two other approaches based on fixed-gaze assumptions: gaze fixating on the task area, and gaze shifted 45 ° towards the window area. The results show that there is a significant difference between luminance distributions driven by gaze and those based on fixed-gaze assumptions, indicating a potentially important impact on glare assessment results as well.