In this study, we experimentally analyze the investment behavior of smallholder farmers in Uganda. We ascertain whether, and to what extent, the real options approach and the classical investment theory can predict farmers’ investment behaviors. We also examine differences in the investment behavior with respect to the presence of a price floor, which is often used to stimulate investments. Furthermore, we look at learning effects. We consider a problem of optimal stopping, stylizing an option to invest in a project. Our results show that both theories do not explain the observed investment behavior exactly. However, our results suggest that the real options models better predict the decision behavior of farmers than classical investment theory. The presence of a price floor has a significant impact on the investment behavior of Ugandan smallholder farmers. Learning from personal experience during the experiment does not have an effect on the investment behavior. However, we find that specific socio-demographic and socio-economic characteristics affect the investment behavior of farmers.