In this paper, we examine conditional, forecast-based dynamic pest management in agricultural crop production given stochastic pest infestations and stochastic climate dynamics throughout the growing season. Using stochastic optimal control we show that correlation between forecast error for climate prediction and forecast error for pest outbreaks can be used to improve pesticide application efficiency. In the general setting, we apply modified Hamiltonian approach to discuss the steady state equilibrium. Given specific functional forms, a closed form solution can be found for the stochastic optimal control problem. Moreover, we find conditions for model parameters so that the optimal pesticide usage path will be monotonically increasing or decreasing in the correlation coefficient between climate forecast errors and pest growth disturbances.