The branch autonomy principle states that the branch carbohydrate economy can be largely independent of the other branches of a tree. This may influence fruit growth and affect global crop yield. While this concept has already been tested on different fruit tree species, branch autonomy has not been characterized with respect to fruit growth in the mango tree. The mango tree, a major fruit production in tropical and subtropical regions, exhibits phenological asynchronisms indicating decorrelated development of branches within a tree and thus a possible autonomy among them. To assess this autonomy, we used a quantitative model of the vegetative and reproductive development of mango tree architecture and fruit quality. This functional-structural plant model combines complementary architectural, phenological and ecophysiological knowledges and relies on two sub-models parameterized for the cultivar Cogshall in Réunion Island. The first sub-model simulates stochastically the development of mango tree architecture, growth units and inflorescences, based on empirical rules. A recent improvement was to take into account leaf mortality to achieve more realistic foliage distribution. Fruit growth and quality development are simulated by a second sub-model that simulates carbon- and water-related processes occurring at the fruiting branch scale during the fruit-growing season. This model assumes the independence of the fruiting branches in terms of carbohydrates synthesis and allocation. We conducted a sensitivity analysis on the size of the fruiting branches and compared the simulated and measured fruit fresh masses at maturity in order to assess the level of autonomy of the branches regarding carbohydrates supply for fruit production. Our results show that the leaf to fruit ratio and the simulated fruit fresh mass increase proportionally with the size of the branches. The comparison with measured fruit masses indicates that a branch size of 2 to 3 GUs are sufficient with re