A farmer’s uncertainty preferences can play a large role in how he makes production decisions on the farm. We attempt to understand how farmers’ household characteristics as well as past harvest shocks affect uncertainty preferences of maize farmers in southern Mexico. By using a series of incentivized lottery games, we estimate coefficients that correspond to Cumulative Prospect Theory, namely the probability weighting function, the curvature of the value function and loss aversion, along with a coefficient for ambiguity aversion. These are estimated controlling for survey data of sociodemographic characteristics as well as maize harvest losses incurred between 2012-2014. Our results provide evidence that having experienced more severe harvest losses leads to more risk aversion and stronger overweighting of small probabilities. Higher losses are not related to loss aversion or ambiguity aversion.