Although there is growing global momentum behind food systems strategies to improve planetary and human health—including nutrition—there is limited evidence of what types of food systems interventions work. Evaluating these types of interventions is challenging due to their complex and dynamic nature and lack of fit with standard evaluation methods. In this article, authors draw on a portfolio of 6 evaluations of food systems interventions in Africa and South Asia that were intended to improve nutrition and identify key methodological challenges and formulate recommendations to improve the quality of such studies. Authors highlight 5 challenges: a lack of evidence base to justify the intervention, the dynamic and multifaceted nature of the interventions, addressing attribution, collecting or accessing accurate and timely data, and defining and measuring appropriate outcomes. In addition to more specific guidance, they identify 6 cross-cutting recommendations, including a need to use multiple and diverse methods and flexible designs. The article also notes that these evaluation challenges present opportunities to develop new methods and highlight several specific needs in this space.