A stochastic dynamic programming model is developed to analyze the integrated control of alfalfa weevil in central New York. The management alternatives considered include biological control by the parasitoidBathyplectes curculionis, cultural control by early harvesting, and chemical control by insecticides. The optimal policies depend upon pest and parasite densities, weather, length of planning horizon, and a number of parameters describing population dynamics. Sensitivity analysis on the parameters indicate that the policies are most sensitive to weather and weevil densities and least sensitive to parasite attack rates. The model results indicate that early harvesting is usually the most effective control procedure for alfalfa weevil management in central New York