Coffee berry borer (CBB) (Hypothenemus hampei; Coleoptera: Curculionidae: Scolytinae) is a major insect pest affecting coffee cultivation that causes large economic losses worldwide. Characteristics related to its life cycle makes it very difficult to control. Usually, CBB control measures are carried out at plot scale, with almost no actions taken at wider landscape scales. It is unclear how plot level control strategies and landscape factors act alone or in combination to influence CBB infestation levels. We evaluated the joint effects of crop management at the plot level, of farm features, and of landscape structure at different spatial scales on CBB infestation in 50 Costa Rican coffee farms. On five plants in each farm, we estimated the maximum number of infested berries during the fruiting period. We measured three separate groups of variables related to plot management practices, farm features and landscape structure. To assess their single and joint contributions, their relative importance and the effects of these variables on the number of infested berries we used the variance partitioning approach of the RandomForest algorithm. When evaluating the groups of factors separately, we found that crop management explained 35% of the variability of number of infested berries, farm features 42% and landscape structure 27%. The joint contribution of all three groups of variables explained 48% of variability of the number of infested berries. However, when we assessed the single contributions of each set of variables, i.e., when controlling the other two set of variables, we found that farm features explained 17% of the variance of the number of infested berries, landscape structure 6% and crop management practices only 3%. The larger amount of the variance explained by the joint effect of crop management practices, farm features, and landscape structure suggests that to develop a pest management strategy at a local scale it is important to consider the effect of bo