Development Finance Institutions have been established to provide finance for long-term economic development in low and middle-income countries. However, these investments are exposed to climate-related risks. This means that Development Finance Institutions need to explore climate adaptation interventions, particularly in view of the fact that developing countries themselves have few resources for climate adaptation. This thesis aims to establish how, using an agent-based model, Development Finance Institutions can achieve long term investments by financing climate adaptation, while creating societal benefits and financial returns. This implies positively contributing to society, for example by creating jobs, generating a stable income, and supporting the resilience of households, while securing profitability for development banks. An investment by the Dutch Development Bank FMO in the cocoa sector in Ghana serves as a case study. Three main steps are followed in the study: (1) translating global climate scenarios at the local level; (2) capturing local behavior in a bottom-up model because understanding the specificity of the local context and individuals’ response is key to really make an impact; and (3) testing climate adaptation interventions. Downscaled climate scenarios are used as input in an agent-based model capturing the behavior of Ghanaian cocoa smallholders. Experimenting with climate adaptation shows that Development Finance Institutions have to make trade-offs in terms of priority, timing, and activity. This step-by-step approach could encourage DFIs to capture the impact of climate change and the dynamics at the local level, and to design tailored interventions. It is also a way to take a holistic perspective and look at the broader picture of economic and societal impacts of climate change, and the response across the whole value chain.