With the rise of the video surveillance system based on cloud computing (hereinafter referred to as the cloud video surveillance system), its complex energy consumption problems, brought by the terminal facilities, physical servers and frequent network- transmissions, can't be ignored. In this paper, the architecture and mechanism of energy consumption optimization of the system are introduced. Then, the energy consumption optimization researches are categorized into three levels: monitoring node, computing node and storage node. Next, considering the existing energy optimization theories and methods applied to the sensor networks and the generalized cloud computing data center, the energy consumption optimization methods for cloud video surveillance system are analyzed and compared with respect to the upper three levels. Finally, several key problems and future research directions for reducing the comprehensive energy consumption of the system are discussed.