A multiple regression model was developed from historical data, 1945-92, to predict pecan [Carya illinoensis (Wangenh.) C. Koch] production in a humid climate. Variables were production trend (year of production), previous year's production, and climatic indices for the previous and current year. Production trend was used to measure change in production with time. Previous year's production was the index of alternate bearing. Variables for previous year's climate were heating degree-days for April-October and cumulative rainfall during May-July and 1-15 Sept. Variables for current year's climate were cumulative rainfall during April-August and 1-15 Sept. The indicator used for scab [Cladosporium caryigenum (Ell. & Langl.) Gottwald] infection was the highest cumulative sum of 2 or more days of consecutive rain occurring in May, June, or 1-15 July. The R2 for the model was 0.908. Production trend was the most important factor influencing production during the 1945-92 study period. Importance of the other variables in decreasing order were previous year's rainfall in May-July, consecutive rainy days, previous year's production, current year's 1-15 Sept. rainfall, previous year's heating degree-days, previous year's rainfall for 1-15 Sept., and current year's rainfall during April-August. Previous year's conditions had a greater effect on production than current year's. The recent decline in pecan production in the southeastern United States is due to an unfavorable change in climate.