The extent and nature of the overlap between visual imagery and visual perception have been debated over the past century. Can visual imagery result in presumably automatic physiological responses such as the pupillary light reflex (PLR)? Laeng & Sulutvedt (2014) reported pupillary responses to dark and bright imagined scenarios. Based on such findings, Kay, Keogh, and Pearson (2022) proposed using the magnitude of imagery-induced PLR as a measure of the ability to generate vivid imagery. We aimed to replicate Kay et al.’s (2022) findings on the PLR response in visual imagery. Ninety-five normally sighted participants were asked to view 16 stimuli in four luminance levels and then imagine the previously seen stimulus. Pupillary responses were measured during both the perception and imagery periods. The PLR response was examined by comparing the pupil diameter in the two darker luminance conditions against that in the two brighter conditions. The PLR response was statistically significant in both the perception and imagery periods (perception: F(1,94) = 598, p < .001; imagery: F(1,94) = 14.7, p < .001). Statistically significant bivariate correlations were consistently observed among the self-report questionnaires (VVIQ, OSIVQ, and SUIS) and the trial-by-trial vividness ratings, suggesting a shared mechanism underlying the subjective evaluation of imagery vividness. However, we could not replicate Kay et al.’s (2022) findings on the association between trial-by-trial vividness ratings and the magnitude of PLR response during the imagery periods. Our results indicate that while pupillary responses may reflect the presence of visual imagery, they do not consistently track the vividness of that imagery. This study highlights the potential of pupillometry as an objective measure of visual imagery but underscores the need for further validation and refinement to improve its reliability and applicability.