Abstract: A kind of fast positioning and harvesting method in oscillation condition for the harvesting robot was researched in order to solve the problem of fruit oscillation influence on harvesting efficiency of harvesting robot. Firstly, it needs to obtain the dynamic and continual images of the oscillatory fruit, then images are recognized and its two-dimensional centroid coordinates are extracted, and then the fruit oscillation period is calculated by FFT algorithm. With measurement of the depth distance of the oscillating fruit, the forward moving speed of the translation joint for the harvesting robot is calculated, and then the harvesting robot starts to harvest, the oscillatory fruit is in balance when it is grabbed. The test results showed that the rate of successful harvesting was 84%, the presented method was better than ever before harvesting method for the oscillating fruit harvest and this method can significantly improve the harvesting speed. In addition, this method is simple and universal, can meet the need of apple and other near-spherical fruit harvesting robot, and the method can provide a reference for realization of practicability and commercialization.