Objective: To investigate whether sleep restriction (SR) worsens working-memory among individuals with depressive and/or anxiety disorders (DepAnx). Method: Ninety college students (17–24 years, 60% female) were administered the structured clinical interview for DSM-IV Axis I Disorders for assessment of DepAnx. They completed a sleep diary and wore actigraphy for 5 days before completing a 2-back spatial working-memory task (2-back), with performance (accuracy and reaction time (RT)) and learning (differences of accuracy and RT across experimental blocks) measures. The average of 5-day sleep duration <6.5 hours was defined as SR. The final sample consisted of 48 healthy controls (26% SR) and 42 with DepAnx (34% SR). Results: The groups were matched on demographics and sleep behaviors (p's > .05). There was a significant interaction effect (SR*DepAnx) on accuracy, F1, 90 = 9.376, p = .003 on 2-back. Post-hoc analysis (LSD) showed that individuals with neither SR nor DepAnx had higher accuracy than other groups (mean differences = .05 to .09, p < .001 to .039). There was a significant main effect of SR on RT, F(1, 90) = 4.116, p = .046. For learning, there was a significant interaction effect on the change of RT from block1 to block2, F(1, 90) = 5.895, p = .017. Post-hoc analysis showed that DepAnx without SR had significantly greater improvement in RT from block1 to block2 than DepAnx with SR (mean difference = 144.9, p = .015). Conclusion(s): The data suggested that sleep restriction was associated with worse executive working-memory, and DepAnx individuals showed further vulnerability on learning aspects of working-memory, warranting clinicians' attention.