The tropical forests are characterized by a great diversity of tree species. Because of the intense human pressures on these resources, many of the species, especially the economically important ones, are rendered either rare, endangered or threatened with incredibly low population sizes. Conventional approaches like ex situ and in situ methods, to conserve the genetic diversity of these populations, are not feasible. In fact in these small and frequently anthropogenically impacted populations, conservation of populations in situ may lead to a loss of diversity through drift and other processes. Under these situations, newer approaches that integrate the genetic variability of the spatially disjunct or fragmented populations should be deployed to help avert the loss of the genetic diversity of the species. We propose an approach based on the establishment of ‘forest gene banks’ wherein the genetic diversity of a species is conserved in situ at a ‘sink’ population into which periodically genes from identified ‘source’ populations are infused. Through such a process, not only is the genetic diversity of the ‘sink’ population enriched and made representative of the others, but also the genetic diversity could be rescued from potentially moribund populations. We discuss the forest gene bank model in the context of the conflicting views on the gene mixing strategies in conservation and argue that under the circumstances of threat to a large number of tropical forest tree species, such approach may have to be followed, albeit with precautions.