We extend the single-stage framework of damage-control inputs like pesticides to a multi-ware framework where four technologies are considered. The economic technologies describe the production and the damages due to pests while the ecological technologies represent the dynamics of pests and predators populations. To account for the possibilities of spatial effects of land uses (crop and non-crop habitats), we consider an analysis at the landscape level and try to find the optimal allocation of the land uses that help minimizing pesticides. To this aim we rely on a prey-predator simulation model. We assess pesticides performance considering nonparametric production frontier techniques. Our results indicate that pesticides can be reduced by 7.7% without reducing the landscape production. In terms of land uses we found that grasslands areas should be increased by more than twice and croplands with medium levels of pesticides unchanged. Croplands with zero and high levels of pesticides should be reduced. In terms of trade-offs between pesticides and the landscape production we found that the spillover effect is very high and result in a negative trade-off because of the destruction of predators by pesticides. Pesticides inefficiency can be reduced when treated areas are spatially aggregated and when grasslands are subsidized. Acknowledgement :