As a consequence of studies indicating that daily plant growth rate can be used to establish specific, numerical, growth‐environment relationships, it has been found that these relationships may be used in the prediction of yields. Design of a prediction system was based on the hypothesis that plant growth rate, as reflected in cumulative growth during the formative portion of plant development, correlated with yield. To insure accessibility of sufficient data, environmental variables were limited to temperature and precipitation because these are generally the only variables published for long periods of time over broad geographical areas. Estimated soil moisture was calculated by the Thornthwaite‐Mather method.Significant lags were found in plant response to environmental variables. Lengths of lag periods also changed during the season. Thus, two prediction equations were used to accommodate these changes and they were applied to the temperature and precipitation records. Resulting cumulative growth values and data representing pre‐season moisture conditions were used as independent variables on which yield was regressed to provide a prediction equation for yield.The system, involving two equations for growth prediction and one equation for yield, was tested on unrelated (data not used in the formulation of equations) weather and yield data for major spring wheat (Triticum aestivum L. em. Thell.) production areas in Canada and the U.S.S.R. The close agreement of predicted yields with actual recorded yields indicates that the use of the system is a promising approach for predicting yields and evaluating the influence of temperature and precipitation on yields.