Termites are major emerging pests of cocoa in Africa. However, their dispersion models on cocoa trees as well as their sampling plans remain unclear. Moreover, the effect of shade trees on the severity of termite infestation on cocoa is poorly known. We evaluated the severity of termite infestation on cocoa trees in relation to shade management in five cocoa agroforestry systems in southern Cameroon with shade cover ranging from rustic to full sun. The dispersion models of termites on cocoa were analyzed using Taylor’s power law, Nachman model and Iwao’s method. We then simulated the optimal sampling sizes for pest monitoring based on parameters of the best-fitted dispersion model using Southwood’s and Green’s formulas. Our results suggest that shade trees reduce belowground infestation by termites and encourage aboveground infestation; however, belowground infestation is more severe. Taylor’s power law was the best-fitted model to describe the aggregative dispersion pattern of termites on cocoa trees. The estimated sampling plans suggested that Green’s sampling was more precise for termites sampling, where less trees is required, than Southwood’s sampling. This study is pioneer on dispersion patterns and sampling plans of termites on cocoa and could be used by researchers and farmers for decision making about pesticide application.