Do agroforestry practices help to improve cocoa sustainability in the humid tropics? Although it sometimes tends to be put under the form of an enthusiastic affirmation, is it not a question that still deserves detailed surveys? This question was initially raised in Côte d'Ivoire by some members of the team who developed a methodology combining questionnaires passed with farmers and direct field observations through density squares. This methodology has been applied to the case of the Amenfi central district of the ghanaian western region, which can be considered as an ¿intermediate¿ region between the old cocoa farms in the East and the young ones in the South-West of the country. In these intermediate zones of the country, agroforestry techniques applied during the 20th century and their mastering by farmers were part of their low risk and low costs strategies. They helped to maintain some plant-diversity in the cocoa farms but did not prevent farmers from massively adopting zero shade techniques and massive deforestation, especially in the extreme western region. As found in Côte d'Ivoire, the old shaded cocoa farms are considered by farmers as a thing of the past, very linked to the old "amelonado" variety that was first introduced. New hybrids sown by farmers, and mostly provided by local research institutions, have been produced under zero shade and high input conditions. The farmers using them tend to reduce shade, even in their old cocoa plots. This paper shows that in the 2000s, farmers need higher returns per hectare and more selective shade to increase the potential of hybrids. This means a much more selective choice of trees than in the past, especially oriented towards trees that have a high value. The idea of rebuilding the old agroforestry pattern is thus a myth. The future of cocoa agroforestry will lie in a commerciallyoriented one, hopefully shored up by some combination with its shade and cocoa protection function. This certainly means a much lowe