Potato sprouting during storage occurs after a break in dormancy, leading to a decrease in quality and consequently economic losses. We used 3379 records from multi-year and multi-environment trials of 537 potato varieties to identify the main factors driving potato dormancy and to develop predictive models for an efficient sprouting forecast. The variety explained the majority of the dormancy variability (60.3%), followed by the year (13.9%) and the location (5.4%). About 250 predictors were considered to develop a predictive model of potato dormancy. The selected model had a validation precision of 14.59 days; it used the variety class and the sum of the daily maximum temperatures in the air during the period from planting to harvest as predictors. The predictions of the selected model were supported by results of the in vivo trial using dormancy measurements from potato varieties grown under different temperature regimes in greenhouse conditions. With the growing impact of climate change on crop production, predictive models as developed here can provide an efficient and costeffective tool to optimize the control of potato sprouting during storage.