In Bangladesh, very few native species have been incorporated into agroforestry systems, and common species are generally exotic pan-tropical species. This neglects the wealth of native genetic resources available and native species are under-valued, which can result in a lack of conservation interest in threatened species. The present study seeks to address this by considering the potential for including a locally threatened native species, Mangifera sylvatica Roxb. into the agroforestry systems of Bangladesh. We investigated farmers’ preferences, tree architecture and root competitiveness in comparison with other agroforestry species to assess the potential and to identify any barriers to use. According to farmers, M. sylvatica is a multipurpose tree species, with the main uses being for timber and provision of wild fruit. In addition, unripe fruits are sold to the pickle industry for seasonal income. The average observed height of matured M. sylvatica trees are approximately 14 m with a straight bole. The crown architecture, light quality under the canopy and root competitive index of M. sylvatica species is similar to popular agroforestry species like Mangifera indica, Artocarpus heterophyllus and Acacia auriculiformis. This indicates M. 9;/indica, A. auriculiformis and A. heterophyllus. Due to over-exploitation, illegal logging and habitat destruction, M. sylvatica is currently threatened in Bangladesh. The major barriers to use include: fruit collection being time consuming and laborious; ripe fruit quality not as good as Mangifera indica; fruits contain small amounts of flesh that often harbour insects; storage problems; irregular fruiting; and limited consumption. Farmer's preferences of wild mango are also simialr to M. indica and A. heterophyllus. Promotion of the species’ positive attributes will be necessary to develop its potential as a new native agroforestry species, which can provide a valuable conservation initiative that ultimately protects genetic diversity and quality of this native tree species.