Biological control of a univoltine race of European corn borer, Ostrinia nubilalis Hübner (Lepidoptera: Crambidae), by Trichogramma in Switzerland is currently timed according to repeated, semi‐field observations of the post‐diapause development of the pest at one site. We developed a phenology model with the aims of replacing this costly labour practice and by considering the Swiss landscape. Observations collected over 29 years were used for model calibration and validation. We parameterized a time‐varying distributed delay model based on published laboratory observations of development durations at constant temperatures. The model was driven with hourly temperature recordings beginning on January 1. The calibration of the mean development rates and their variations was based on semi‐field data of larval and pupal development. The model, with its calibrated parameters and their variations, allowed the simulation of mean predictions and prediction intervals. We validated the model predictions of emergence with flight data (obtained via light traps) from several sites in western and northern Switzerland. The simulated mean emergence was 6.9 degree‐days earlier than the observed flight at the calibration sites and only 0.5 degree‐days earlier than the observed flight at other sites. Our simulation of pupation explained half of the variation in emergence time, whereas semi‐field observations of pupation explained three‐quarters of this variation. Our model simulations are not subjected to the local potential biases. Simulation errors from a year with an extremely dry spring were explained by the lack of consideration of humidity by the model. Our simulations provide a valid and less labour‐intensive alternative to observations for timing biological control in the maize‐growing areas of Switzerland and likely other areas of Europe.