Complex social problems have multilevel roots and intricate webs of causation. They arise at the intersection of several of the spheres that comprise modern life: the physical, biological, cultural, political, social, economic, technological and environmental. They challenge the limited cognitive capacities of human beings, as delays, nonlinearities, policy resistance and complex interactions between the different parts of a system produce unforeseen consequences and patterns of systemic behaviors that are difficult to anticipate. Incentivized by the short-term focus of the political system, policy makers tend to act on the most visible parts of problems, which are often not causes but symptoms. They produce policies that often lead to worse conditions over time. Learning in complex systems, on the other hand, is hindered, because feedback from policies is typically opaque and amenable to distortion. This study draws on the complexity sciences, system dynamics, social marketing and related fields to address the issue of human development. We first explore the characteristics of complex social problems, showing that disciplines such as social marketing, based on individualistic mindware, are ill equipped to address them. The dissertation defends the need for a new interdisciplinary field (complex social problems management) capable of dealing with the challenging social problems of our times. In particular, we focus on the replication of low human development across generations, discussing the concepts of poverty, well-being and human development as well as the conditions that foster the development of essential psychobiological mechanisms in the first years of life. The discussion encompasses the social ecosystems that constrain opportunities for the poor and reinforce detrimental structures that produce toxic stress and low attention to human capital. A broad causal loop diagram summarizes the feedback loops that explain the persistence of this phenomenon. Based on the discussion in the first three chapters, we present three system dynamics models in the subsequent chapters, each one addressing an aspect of the problem. The first model, which we call the Caped model, focuses on parents' and children's essential capabilities and on the social environmental demands that deplete parents' psychological resources. The second model broadens the scope of the first and focuses on explaining why societies (in particular, Brazil) that are dominated by rent-seeking and extractive institutions produce detrimental social ecosystems. We uncover what seems to be a central mechanism, which is narrative (mindware) control. The third model deals with social change in general, and it aims to describe the channels and conditions by which that change occurs. It complements the analysis in the preceding chapters by identifying how a deep societal change toward the promotion of human development can occur. This dissertation contributes to the study of human development by employing a method that is suitable for modeling complex social...