In the present study, the effect of weather on maize yields in northern China was examined using data from 10 districts in Inner Mongolia and two in Shaanxi province. A regression model with a flexible functional form was specified on the basis of agronomic considerations. Explanatory variables included in the model were seasonal growing degree days, precipitation, technological change (e.g. adoption of new crop varieties, improved equipment, better management, etc.) and dummy variables to account for regional fixed effects. Results indicated that a fractional polynomial model in growing degree days could explain variability in maize yields better than a linear or quadratic model. Growing degree days, precipitation in July, August and September, and technological changes were important determinants of maize yields. The results could be used to predict potential maize yields under future climate change scenarios, to construct financial weather products and for policy makers to incentivize technological changes and construction of infrastructure (e.g. irrigation works) that facilitate adaptation to climate change in the agricultural sector.