To solve the problems of the existing orchard apple post-harvest transportation equipment with a single autonomous navigation mode and the inability to start or stop at any point, a dual navigation mode small orchard transportation robot was developed and designed, whose hardware system mainly included five modules, such as pedestrian led navigation, fixed-point navigation, control, motion and power, and the software system included three modules, such as command interaction layer, information processing layer and execution layer. The user selected the navigation mode under the command interaction layer according to the demand. In the selected navigation mode, the information processing layer processed the target point pose information provided by the pedestrian led navigation or fixed-point navigation module and then used OpenPose human pose recognition-based target tracking control or Real time kinematic-global navigation satellite system（RTK-GNSS）based distance-angle control methods to output the next moment of motion velocity information to the execution layer. And then the motion and power module realized the pedestrian led navigation and fixed-point navigation in the orchard environment based on the velocity information. When the transport robot operated in the orchard with a rated load of 200kg and speed of 0.5m/s, the average value of target tracking error was less than 9cm and the standard deviation was less than 4cm in the pedestrian led navigation mode;the relative error of reaching the target point was less than 13cm and the standard deviation was less than 1.5cm in the fixed-point navigation mode, and the absolute error was less than 7cm and the standard deviation was less than 0.5cm. The test results showed that the robot can meet the requirements of autonomous transportation and safe obstacle avoidance in orchards.