As agricultural hilly watersheds are widespread throughout the world, there is a strong need for reliable estimates of land surface fluxes, especially evapotranspiration, over crop fields on hilly slopes. In order to obtain reliable estimates from eddy covariance (EC) measurements in such conditions, the current study aimed at proposing adequate planar fit tilt corrections that account for the combined effects of topography, wind direction, and vegetation height on airflow inclinations. EC measurements were collected within an agricultural hilly watershed in northeastern Tunisia, throughout the growth cycles of cereals, legumes, and pasture. The wind had two dominant directions that induced upslope and downslope winds. For upslope winds, the airflows were parallel to the slopes and slightly came closer to the horizontal plane when vegetation grew. For downslope winds, over fields located in the lee of the rim top, the airflows were almost horizontal over bare soil and came closer to the topographical slope when vegetation grew. We therefore adjusted the planar fit tilt correction on EC measurements according to vegetation height and by discriminating between upslope and downslope winds. This adjusted tilt correction improved the energy balance closure in most cases, and the obtained energy balance closures were similar to that reported in the literature for flat conditions. We conclude that EC data collected within crop fields on hilly slopes can be used for monitoring land surface fluxes, provided planar fit tilt corrections are applied in an appropriate manner.