Governments around the world seek to create programs that will support sustainable agriculture and achieve food security, yet they are faced with uncertainty, system complexity and data scarcity when making such choices. We propose decision modeling as an innovative approach to help meet these challenges and offer a case study to show the effectiveness of the tool. We use decision analysis tools to model the possible nutrition-related outcomes of the Ugandan government’s long term agricultural development plan termed ‘Vision 2040’. The analysis indicates potential shifts in household nutritional contributions through the comparison of the current small-scale diverse systems and the envisioned industrial agricultural systems that may replace them. A Monte Carlo simulation revealed that Vision 2040 plans outperform homegardens in terms of energy and some macronutrients, yet homegardens are likely to be better at producing key vitamins and micronutrients, such as Vitamin A. Value of information calculations applied to Monte Carlo outputs further revealed that gathering more data on the annual yields and nutrient contents of staples, pulses, vegetables, and fruits could improve certainty about the nutrition contribution of both scenarios. We conclude that the development of Uganda’s agricultural sector should consider the role that agrobiodiversity in the current small-scale agricultural systems plays in national food and nutrition security. Any changes according to Vision 2040 should also include farmers' voices and current crop management systems as guides for a sustainable food supply in the region. This modeling approach may be a tool for governments to consider agricultural policy implications, especially given the data scarcity and agricultural variability in regions such as East Africa.