While the production effect of El Niño and its counterpart – La Niña – is well documented, many of the previous studies apply binary variables (e.g., for El Niño and La Niña events) to analyze the relationship, and, moreover, much of their conclusions rely on in-sample fit and test statistics. In this study, we extend previous literature by examining the effect of El Niño Southern Oscillation (ENSO) on U.S. corn production in an out-of-sample setting. In so doing, we incorporate weather variables, such as degree days and precipitation in the analysis, to investigate the trade-off between model uncertainty and parameter uncertainty. We find that ENSO likely impacts U.S. corn production though extreme degree days. This is particularly true for Counties in the southern Corn Belt as well as corn-growing Counties in the Appalachian and Southeastern regions. In many instances, however, more accurate forecasts are obtained when the ENSO effect on corn yields is modeled directly.