Angiosperm phylogenetics investigates the evolutionary history and relationships of angiosperms based on the construction of phylogenetic trees. Since the 1990s, nucleotide or amino acid sequences have been widely used for this and angiosperm phylogenetic analysis has advanced from using single or a combination of a few organellar genes to whole plastid genome sequences, resulting in the widely accepted modern molecular systematics of angiosperms. The current framework of the angiosperm phylogeny includes highly supported basal angiosperm relationships, five major clades (eudicots, monocots, magnoliids, Chloranthales, and Ceratophyllales), orders grouped within these clades, and core groups in the monocots or eudicots. However, organellar genes have some limitations; these involve uniparental inheritance in most instances and a relatively low percentage of phylogenetic informative sites. Thus, they are unable to resolve some relationships even when whole plastid genome sequences are used. Therefore, the utility of biparentally inherited nuclear genes with more information about evolutionary history, has gradually received more attention. Nevertheless, there are still some plant groups that are difficult to place in the angiosperm phylogeny, such as those involving the relative positions of the five major groups as well as those of several orders of eudicots. In this review, we discuss the applications, advantages and disadvantages of marker genes, the deep relationships that have been resolved in angiosperm phylogeny, groups with uncertain positions, and the challenges that remain in resolving an accurate phylogeny for angiosperms.