Annual jack-pine budworm,Choristoneura pinus pinusFreeman, population samples and weather records from Douglas County, Wisconsin during 1959–1976 were used to formulate prediction models using forward (stepwise) inclusion multiple regression analyses. Final models for late larval and pupal counts and defoliation ratings consisted of temperature precipitation, and population terms at years t-1 and t-2. The accepted model for the late larval counts explained 64% of the variation in the observed data, the model for pupal counts explained 70%, and both models indicated an association of weather conditions with larvae overwintering in hibernaculae. While the defoliation model had an adjusted R2of 62%, abnormalities were present in the residuals suggesting caution for use in prediction. Validation showed that the models mimicked observed population and defoliation trends and hence, could be helpful in predicting future budworm outbreaks in susceptible jack pine stands.