AbstractSahlbergella singularis Haglund and Distantiella theobroma (Distant) (Heteroptera: Miridae) are the key insect pests of cacao in Ghana and elsewhere in West Africa. Since 1954, spraying with synthetic insecticides has been the principal means of controlling these pests. In West Africa, environmental concerns, economic stimuli, and unreliable access to insecticides have stimulated interest in ecologically more benign mirid control strategies as an alternative to a total reliance on insecticides. Males of both mirid species, and those of the less damaging Bryocoropsis laticollis Schumacher, respond to the same synthetic sex pheromone blend, so pheromone‐based strategies may provide control as well as monitoring opportunities. Pheromone traps were deployed for 3 months at nine densities between two and 30 traps per 0.1 ha plot (20–300 traps ha−1) plus an untreated control treatment in a replicated large‐scale field experiment on mature mixed Upper Amazon hybrid cacao in Ghana, in order to determine the optimal dispenser density for mass trapping, lure‐and‐kill, and/or lure‐and‐infect. At the end of the trapping period, mirid populations in the various treatments were assessed by insecticide knockdown on 400 trees and by searches to hand height on 1 200 trees, together with an assessment of mirid damage. In total 781 S. singularis and 235 D. theobroma were captured in the pheromone traps. The optimal dispenser density for S. singularis was 150 traps ha−1. Over 300 traps ha−1 was probably optimal for D. theobroma as a smaller proportion of the population was captured, and numbers caught per trap displayed no decline with increasing trap density. From insecticide knockdowns, mirid population density was estimated at 220–230 ha−1, 63% of which were D. theobroma. Too few pods and orthotropic shoots were damaged by mirids to establish the efficacy of pheromone trapping for mirid control.