It is generally assumed that compounds are emitted from flowers in order to attract and guide pollinators. Due to the invisibility and the highly variable nature of floral scent, no efficient and reliable methods to screen for genetic variation have been developed. Moreover, no convenient plant model systems are available for flower scent studies. In the past decade, several floral fragrance-related genes have been cloned; the biosynthesis and metabolic engineering of floral volatiles have been studied with the development of biotechnology. This review summarizes the reported floral fragrance-related genes and the biosynthesis of floral scent compounds, introduces the origin of new modification enzymes for flower scent, compares different methods for floral fragrance-related gene cloning, and discusses the metabolic engineering of floral scent. Finally, the perspectives and prospects of research on floral fragrance are presented.