The aim of this paper is to present an original, integrated approach developed in Reunion Island by Cirad in collaboration with other agronomic research organizations (Inra-France, Csiro-Australia). Our objective is to improve our understanding of the mango tree architecture and physiology in relation to fruit quality and shelf life in order to improve orchard management techniques (pruning, irrigation, harvest date...) and post-harvest practices. We postulate that innovative approaches should restrict or even ban the use of chemicals in order to meet the requirements of consumers for integrated fruit production. Our central hypotheses are that: i) tree functioning is mainly controlled by the availability of carbohydrates, ii) flower induction is controlled by both carbohydrate status and leaf age/temperature, and iii) fruit quality depends on pre-harvest (tree carbohydrate status) as much as on post-harvest conditions. The actual stage of development of our approach is presented. Our major validated models are a biochemical model of leaf photosynthesis capable to account for the presence of inflorescences and fruits, and a carbon and water-based model of fruit quality determination at the branch level. Models of carbohydrate management at tree level and phenology are in the process of being built. The possibility to include the effect of irrigation on photosynthesis and the effect of growing conditions on the post-harvest behaviour of mangoes are currently investigated. Additional studies have been carried out with the aim of extrapolating our findings to cultivars other than ¿Cogshall¿, our major plant model so far. The expected outcomes of our modelling approach are discussed in the light of our results. The prospect of including pest and disease management in our global approach is evoked.