In traditional orchards, lots of labors and material resources are required to carry fruits, spray pesticides and remove weeds. It has become increasingly prominent that autonomous labor burdens should be reduced. In this paper, a general orchard platform based on Field Programmable Gate Array (FPGA) is designed, which can be equipped with mowing and spraying systems to complete the whole orchard path. However, due to the limitation of orchard circumstances and other factors such as extreme illumination or light interference, there will be some position deviations in navigation by machine vision. To solve this problem, a method of orchard path recognition and location based on machine vision is proposed. The results show that this method can effectively adjust the deviation of path recognition through cameras and the identification and positioning accuracy are improved of the general orchard platform.