Over the past several years, increased rates of mortgage foreclosures in the United States have had significant impacts on economies around the world. As part of the societal response to this problem, nonprofit community development corporations acquire foreclosed properties in troubled neighborhoods and redevelop them to maintain stability. This paper addresses a decision problem faced by these organizations on what types of properties to acquire given limited available resources. We develop dynamic and stochastic programming models to assist such organizations in choosing foreclosed properties for acquisition and redevelopment, taking into account uncertain market conditions that change over time. We first analytically study a dynamic programming model under some special cases, and then seek insights for the general stochastic model through a multistage stochastic programming formulation and its numerical analysis. Although our analysis is based on a problem in the United States with potential implications for the global economy, our research is applicable to non‐U.S. government actors at the local and regional level as well.