The yield gap is the difference between the potential and actual yield. Knowledge of its size and spatial distribution is useful for technology targeting, to optimize input and provide field specific management recommendations to the farmers.  Our study area covered more than 500 km2 in the Frailesca, in the state of Chiapas, Mexico. Potential yield of maize under rain-fed, i.e., potentially water limiting conditions was simulated with the SALUS crop simulation model. A RapidEye satellilte image taken around tasseling was used to identify maize fields and to estimate ground cover. Based on yield data obtained from farmers’ fields, a regression of ground cover vs. yield was calculated to predict actual yield of all maize fields in the study area. Information on potential and actual yield, as well as detailed records on crop management from several hundred fields were subsequently fed into a data mining tool (WEKA) to derive field specific, optimized crop management recommendations for maize.