Studies have revealed that information can be intentionally forgotten when instructed, commonly studied in the laboratory with the directed forgetting (DF) procedure. The current investigation examined pupillometric signals associated with intentional forgetting, as the pupil reflects the activity in the locus coeruleus-norepinephrine (LC-NE) system that is functionally involved in the neural correlates of intentional forgetting. Experiment 1 employed an item-method DF paradigm, where participants were presented with natural scenes, each followed by a memory cue to either remember (R) or forget (F) that scene. At test, participants were asked to judge whether the presented scene was the original studied version (i.e., "Old") or a mirrored variant (i.e., "Lure"). By comparing pupil dilation during test trials between R-cued and F-cued scenes for both hit and miss trials, we found greater pupil dilation for F-cued miss trials compared to R-cued miss trials, but no difference in pupil dilation between the cue conditions for hit trials. This suggests a unique pupillometric pattern linked to successful intentional forgetting. Experiment 2 was aimed at assessing if memory strength differences could provide an explanation for the observed effect. Instead of DF cues, we manipulated memory strength by repeating a subset of scenes, thereby converting all study items into R-cued items with different degrees of familiarity. We observed no difference in pupil dilation between strongly encoded and weakly encoded scenes at test, indicating that encoding strength by itself did not explain the difference in pupil dilation resulting from intentional forgetting. Together, these findings provide novel evidence that pupil fluctuations during retrieval index successful intentional forgetting.