Estimates of daily lake evaporation based on energy‐budget data are poor because of large errors associated with quantifying change in lake heat storage over periods of less than about 10 days. Energy‐budget evaporation was determined during approximately biweekly periods at a northern Minnesota, USA, lake for 5 years. Various combinations of shortwave radiation, air temperature, wind speed, lake‐surface temperature, and vapour‐pressure difference were related to energy‐budget evaporation using linear‐regression models in an effort to determine daily evaporation without requiring the heat‐storage term. The model that combined the product of shortwave radiation and air temperature with the product of vapour‐pressure difference and wind speed provided the second best fit based on statistics but provided the best daily data based on comparisons with evaporation determined with the eddy‐covariance method. Best‐model daily values ranged from −0.6 to 7.1 mm/day over a 5‐year period. Daily averages of best‐model evaporation and eddy‐covariance evaporation were nearly identical for all 28 days of comparisons with a standard deviation of the differences between the two methods of 0.68 mm/day. Best‐model daily evaporation also was compared with two other evaporation models, Jensen–Haise and a mass‐transfer model. Best‐model daily values were substantially improved relative to Jensen–Haise and mass‐transfer values when daily values were summed over biweekly energy‐budget periods for comparison with energy‐budget results.