Adverse social experiences are often linked to suicidal behavior in adolescence, perhaps particularly for girls. Social problem-solving abilities may indicate more or less adaptive responses to adverse social experiences that contribute to adolescent girls' risk for suicidal behavior. While social problem-solving is implicated in cognitive and behavioral theories of suicidal behavior, prior work is largely cross-sectional and examines bivariate associations between social problem-solving, assessed in neutral conditions, and suicidal behavior. Using a novel performance-based task, this study assessed social problem-solving in adolescent girls (N = 185, Mage = 14.66, SD = 1.41) before and after an experimentally simulated social stressor and examined associations between social problem-solving and past-year suicidal behavior. Prospective analyses tested whether greater changes in specific social problem-solving domains following the social stressor predicted greater likelihood of suicidal behavior over a 9-month follow-up in contexts of elevated, real-life interpersonal stress. Results revealed that adolescent girls who showed greater changes (i.e., reflecting declines) in problem-solving effectiveness following acute social stress were more likely to exhibit suicidal behavior over the following 9 months, but only if they also experienced elevated interpersonal stress in real life. State-dependent changes in social problem-solving may indicate a cognitive vulnerability following social stress that, in combination with cumulative interpersonal stress in real life, distinguishes adolescent girls at heightened risk for future suicidal behavior. Findings demonstrate the importance of examining suicide risk factors under conditions that may more closely mirror the interpersonal contexts in which adolescents' risk for suicidal behavior may be elevated.