This paper reviews and evaluates a set of studies that utilize dynamic systems (DS) principles, and in two cases dynamic systems methods, for predicting antisocial development and other behavioral outcomes. I suggest that the emphasis of DS approaches on process and nonlinear causation is very different from the emphasis in developmental psychopathology on prediction, yet the marriage of these approaches is necessary to capture the complex interactions that give rise to problematic trajectories. The studies reviewed do indeed uncover predictive relations that would have been difficult to conceptualize or impossible to find using more traditional strategies. In discussing these studies, I suggest DS interpretations of emerging individual differences, phase-specific change, sleeper effects, mediating variables, and behavioral rigidity versus malleability, in the context of developmental prediction. I also discuss the advantages of moving from DS concepts to DS methods in clinical-developmental research.