Both pain and negative affect (e.g., pain catastrophizing) interrupt and demand attentional resources, which implicates working memory (WM)—the short-term maintenance and manipulation of information. While these factors may not lead to observable WM performance deficits on easy cognitive tasks, an increase in cognitive load (difficulty) may make it more difficult to conserve accuracy. However, few studies have examined pain, pain catastrophizing, and cognitive load concurrently to evaluate their interactive effect on WM. 42 healthy undergraduate students aged 18-22 completed a WM task (n-back) under randomly ordered pain and no-pain thermal stimulation conditions. Participants responded when the current stimulus matched the stimuli presented n positions back in the sequence, with load increasing from n=1 to n=3. Performance was evaluated using hits (correct responses) and false alarms (incorrect responses). Participants also completed the Pain Catastrophizing Scale. Linear mixed effect models were employed to evaluate the interactive effect of pain catastrophizing, experimental pain, and cognitive load on WM performance. We found a significant three-way interaction among pain catastrophizing, pain condition, and cognitive load on false alarms (b = -.12, SE = .047, t = -2.54, p = .012), 95% CI [-.21, -.029]. Under low load, low catastrophizers had fewer false alarms than high catastrophizers under both pain conditions. However, under high load, low catastrophizers had fewer false alarms than high catastrophizers when pain was on, but more false alarms than high catastrophizers when pain was off. Our findings suggest a complex relationship between pain and WM performance, dependent on task difficulty and individual differences in pain catastrophizing. Results suggest that while catastrophizing alone may not consume enough resources to significantly affect accuracy, even under high load, the combination of pain and catastrophizing overwhelms the increased effort needed to maintain performance under high load, leading to decreased accuracy.