Background and Aims Rhododendron is a species-rich and taxonomically challenging genus due to recent adaptive radiation and frequent hybridization. A well resolved phylogenetic tree is conducive to understanding the diverse history of Rhododendron in the Himalaya-Hengduan Mountains where the genus is most diverse. Methods We reconstructed the phylogeny based on plastid genomes with broad taxon sampling, covering 161 species representing all eight subgenera and all 12 sections, including approximate 45% of the Rhododendron species native to the Himalaya-Hengduan Mountains. We compared this phylogeny with nuclear phylogenies to elucidate reticulate evolution events and clarify relationships at all levels within the genus. We also estimated the timing and diversification history of Rhododendron, especially the two species-rich subgenera Rhododendron and Hymenanthes that comprise >90% of Rhododendron species in the Himalaya-Hengduan Mountains. Key Results The full plastid dataset produced a well resolved and supported phylogeny of Rhododendron. We identified 13 clades that were almost always monophyletic across all published phylogenies. The conflicts between nuclear and plastid phylogenies strongly suggested that reticulation events may have occurred in the deep lineage history of the genus. Within Rhododendron, subgenus Therorhodion diverged first at 56 Mya, then a burst of diversification occurred from 23.8 to 17.6 Mya, generating 10 lineages of the component 12 clades of core Rhododendron. Diversification in subgenus Rhododendron accelerated c. 16.6 Mya and then became fairly continuous. Conversely, Hymenanthes diversification was slow at first, then accelerated very rapidly around 5 Mya. In the Himalaya-Hengduan Mountains, subgenus Rhododendron contained one major clade adapted to high altitudes and another to low altitudes, whereas most clades in Hymenanthes contained both low- and high-altitude species, indicating greater ecological plasticity during its diversification. Conclusions The 13 clades proposed here may help identify specific ancient hybridisation events. This study will aid to establish a stable and reliable taxonomic framework for Rhododendron, and help to provide insight into what drove its diversification and ecological adaption. Denser sampling of taxa, examining both organelle and nuclear genomes, is needed to better understand the divergence and diversification history of Rhododendron.