Plant regeneration through in vitro techniques is commercially achieved for a large number of annual species and some perennials. Among perennials, rubber tree is one of the most recalcitrantspecies and clonal planting material is still restricted to buddedscion clones on rootstocks derived from seedlings. This talk aims to overview development of somatic embryogenesis in Heveabrasiliensis and discuss recent advances and limitations. Somatic embryogenesis is the most suitable technique for large-scale plant multiplication of clones at low cost of production. Several processes were attempted in Hevea. Primary somatic embryogenesis generated low quantity of in vitro plants but for more than twenty Hevea clones (Montoro et al. 2012). Field tests revealed a higher growth and latex production for this plantig material compared with conventional budded clones (Carron et al. 2009). Both absence of graft and rejuvenation process were expected to play a role in the superiority of this material. For a few number of clones, microcutting allowed getting a large quantity of plants from a few in vitro plants regenerated by primary somatic embryogenesis (Hua et al. 2010). Long-term somatic embryogenesis processes were attempted in order to develop large-scale propagation method including several ways of secondary somatic embryogenesis (Lardet et al. 2009). These processes have two main limitations. First, they are restricted to a few numbers of clones. Second, long-term multiplication of embryogenic friable calliincreases risk to induce somaclonal variation. To date, no strict somaclonal variant was described in rubber tree. In some process, the use of cryopreservation and attention to culture conditions are proposed to reduce the risk of somaclonal variation (Lardet et al. 2006). Direct secondary somatic embryogenesis without passing through the stage of callus was recently developed. The direct production of secondary embryos from primary embryos was obtained for about seven c