In recent years the quality of the speech generated by text-to-speech synthesis has been improved dramatically. However the naturalness of the synthesized sound can be further improved to have more emotions to imitate human kind speaking. This inspires the research of expressive speech synthesis (ESS) to improve naturalness. Currently, there have been a lot of papers focusing on the development of ESS based text-to-speech systems. This study tries to give an overview of ESS studies and summarize five research topics-speaking styles, emotion categories, corpus construction, ESS approaches, and the evaluation of synthetic sounds.