The objective of this work was to develop a simple model for predicting water use for the potato crop in Maine using daily maximum and minimum air temperatures, and rainfall as model inputs. A reasonably accurate model would be helpful for growers who irrigate to judge when to apply water to their crop. The model developed uses historic data on average daily water use by potatoes (reported for the Aroostook Research Farm, in Presque Isle, Maine) as an initial basis for predicting crop water use on a given day. The historic value of crop water use is then proportionally raised or lowered based on: (1) mean temperature for the day versus the historic average temperature for that day; (2) the range between maximum and minimum temperature versus the historic average range in temperature. For use with potatoes in Maine, the model assumes that the season starts with a full profile and 2.5 inches of available water at field capacity to a depth of 20 inches. Rainfall is added, and estimated crop water use is subtracted, from the soil water balance on a daily basis. To evaluate the model, observed soil water balance (estimated from soil water potential measurements) was plotted against predicted soil water balance for the drought year of 2002. Across the whole range of soil water balance in the drying cycle, the r² of predicted versus observed values was 0.50. In the range where most irrigation decisions would be made (> 1.5 inches of available soil water), the model gave an r² of 0.90. The growers who have used the model in pilot studies have judged it a useful tool.