To characterize the population dynamics of Erionota thrax Linnaeus and its major parasitoids, biweekly samplings were carried out in a commercial Cavendish banana plantation and a subsistence farm of local varieties, from April 2004 to November 2005. Five primary endoparasitoids were recorded: Ooencyrtus erionotae Ferriere, Cotesia erionotae Wilkinson, Brachymeria albotibialis Hoffmann, Elasmus sp. and Melaloncha sp. Ooencyrtus erionotaeC. erionotae, and B. albotibialis were, respectively, the major egg, larval and pupal parasitoids. In 2004, there was no significant difference in the E. thrax population densities and parasitism of immature life stages in both farms. In 2005, the pest population densities were significantly different, while only egg and larvae parasitism rates were significantly different. In the commercial plantation, peaks of E. thrax population densities per banana mat were recorded in April, June, July, and October–December and highest values for eggs, larvae and pupae were 4.66, 2.76 and 1.7, respectively. Percentage parasitism peaked in August, October, and February with peak values ranging from 60 to 100%. In the subsistence farm, peak E. thrax population densities were observed between May–July and September–December. Percentage parasitism peaked in January, February, July, August, October and November, with peak values ranging from 24 to 100%. Correlations between population densities of E. thrax and the numbers of parasitized insects showed strong positive relationships. There was evidence of a delayed density dependence of the parasitoids as most peaks of parasitism appeared 1 or 2 months after those of the host stages. Parasitism rates in the commercial plantation were substantial in spite of the frequent use of insecticides, and activities of the parasitoids were more favoured in the subsistence farm compared with the commercial plantation.