Reports of emotions experienced over the past week can be influenced by memory bias, which is more pronounced for people with depression. No studies, however, have examined memory bias for specific emotion clusters (e.g., sadness, anxiety, and anger) experienced on a day-to-day basis among people with depression or a history of depression. Participants (N = 1,657) from the Midlife in the United States Study were assessed for depression. Approximately 6 months later, participants reported their emotional experiences for 8 days and recalled these experiences on the final day. Differences in recalled and reported emotion were compared between participants with and without a history of depression. Participants overestimated experience only of negative emotions, particularly anger, and this negativity bias was greatest for participants with a history of depression. Feelings related to anger were prone to greater overestimation than sadness or anxiety. These findings emphasize the role of memory bias in retrospective reports of specific emotions and illustrate the presence of an amplified memory bias among people who are at a greater risk for recurrent depressive episodes.