Research and development of fruit-picking robot is of great significance for improving harvest efficiency, guaranteeing fruit quality and reducing labor intensity. However, picking-robot under the effect of non-structural factors including over lapping, adhesion and occlusion between bunched fruits may damage the fruits very easily due to inaccurate target positioning, incorrect clamping-cutting sequence and improper clamping-cutting gesture. The main cause of the damage is that the coupling mechanism between visual cognition and clamping-cutting actuators for undamaged picking has not been well settled. In order to comb the latest research progress on anti damage operation of fruit-picking robot, this paper makes a comprehensive review and analysis on undamaged harvesting from three aspects: the multi-dimensional information of targets by visual perception, such as picking point, peduncle position, bounding volume of grape for undamaged fruit-picking and so on; visual cognition and intelligent anti damage picking planning of fruit-picking robots; anti damage picking mechanism design and its behavior control. Moreover, the key problems that need to be solved in the future are summarized and prospected. This paper may provide reference and views for further research on the problems of fruit intelligent undamaged picking in unstructured environment.