Many researchers have used the standard Iowa Gambling Task (IGT) to assess decision-making in adolescence given increased risk-taking during this developmental period. Most studies are cross-sectional and do not observe behavioral trajectories over time, limiting interpretation. This longitudinal study investigated healthy adolescents' and young adults' IGT performance across a 10-year span. A total of 189 individuals (aged 9-23 at baseline) completed a baseline session and were followed at 2-year intervals yielding 5 time-points. IGT deck contingencies were shuffled over time to reduce practice effects. IGT performance (good minus bad decisions) was measured at each assessment point and separated into 3 metrics: overall performance (all blocks), decision-making under ambiguity (blocks 1 and 2), and decision-making under risk (blocks 3, 4, and 5). Covariates included estimated intelligence and affective dispositions as measured by the Behavioral Inhibition and Activation System (BIS/BAS) Scales. A linear effect of age yielded the best fit when comparing linear and quadratic effects of age on overall IGT performance. Age and intelligence positively predicted overall performance, whereas affective approach tendencies (BAS) negatively predicted overall performance. Practice effects were observed and controlled for. Models of ambiguity and risk metrics yielded different patterns of significant predictors. Age predicted better performance and affective approach tendencies predicted worse performance for both metrics. Intelligence was a significant predictor for risk, but not ambiguity. This longitudinal study extends prior work by showing age-related improvements in reward-based decision-making and associating those improvements with cognitive and affective variables. Implications of the results for adolescent development are discussed.