Rubber breeding programme at the Rubber Research Institute of Malaysia (RRIM) was started in 1928. Since then, the Institute has gone through six phases and produced six series of RRIM clones ranging from RRIM 500 to the more recently selected set of clones which have been provisionally given RRIM 1000 series numbers. Hevea breeding in the early days was devoted to mother tree selection and empirical crosses. The results of this successful work is amply reflected in the statistics of rubber planting up to the end of 1992 which shows the planting of these old clones in some of our rubber plantations even till today. Later Hevea breeding proceeded to the conventional breeding system of cyclical generation-wise assortative mating (GAM), where the best genotypes in one generation were used as parents for the next cycle of breeding. This approach has given rise to the present generation of clones whose experimental yield perrformances are given in this paper. With the recent interest shown in rubberwood by the wood-based industries for the manufacturing of furniture and its components, emphasis is now being shifted towards breeding and selection of quick growing clones which show potential for latex yield as well as timber. Currently, selected RRIM 900 series clones together with selected clones of the provisionally named RRIM 1000 series have shown promise for production of both these products. Further Hevea improvement through plant breeding is possible and necessary to sustain the well-being of the Natural Rubber industry. This paper discusses some aspects of future breeding strategies in the RRIM breeding programme. Among the strategies discussed are (1) broadening the genetic base of rubber, (2) breeding for resistance to leaf diseases by using prospective parental materials from among the 1981 germplasm collection, (3) planning crosses to improve further yield, girth and resistance to leaf diseases, and (4) researching for an early and reliable method of selection for yield and timber.