With the rapid development of modern agricultural technology, plant factory has become the most advanced development stage of facility agricultural. At present, the majority of work tasks in plant factory completed by manpower are labor-intensive and low efficient, therefore, the agricultural intelligent equipment system has become a hot spot in the development of plant factory. In view of the task demand of the carrying and spraying of the three-dimesional seedling tray, the three-dimesional seedling tray management robot was developed. In order to make the manipulator of three-dimesional seedling tray management robot complete all carrying and spraying tasks flexibly and efficiently, meanwhile to reduce operating space and structure size of manipulator, parameters of the manipulator were optimized by the method of theory and experiment. Firstly, in order to determine the relationship between the end coordinate of the manipulator’s connecting rod and the base coordinate system, the kinematic model of the robot system was established by D-H method, which was important theoretical basis for the workspace analysis. Then the workspace of manipulator was constructed by graphic method, and the workspace constraint conditions were determined according to the condition that manipulator workspace accommodated target workspace. Based on that, the objective function was established according to shortest distance and compact structure, and genetic algorithm was used to solve the objective function. The optimal rod lengths (big arm, medium arm, small arm) of the manipulator were 648, 472, and 396 mm, and the limit values of the optimal joint angle were 96○, 68○, and 126○. The workspace and the target workspace of the robot were depicted in the MATLAB (Matrix Laboratory) software platform according to the optimal solution of the manipulator parameters, the kinematics equation of the robot and the range of the manipulator’s parameters. The simulation result showed that the target workspace was between the inner limiting envelope interfaceand the outer limiting envelope interface of the manipulator, which verified the manipulator’s ability to cover the target workspace, and the rationality of the theoretical optimization for the parameters of the manipulator was proved. Finally, in order to further validate whether the manipulator could complete all the action tasks of the target workspace, the robot prototype and the three-dimesional seedling tray experimental platform were built in the laboratory, and the motion planning test of carrying and spraying of the robot system prototype was carried out. The carrying test was planned as follows: According to the target workspace size and the theoretical position coordinate value, the manipulator was controlled to move vertically upward from the lowermost (lower limit) to the topmost (upper limit) of the target workspace, this group of actions were repeated 100 times, and seedling tray was always placed horizontally during carrying. The carrying test mainly verified the manipulator’s ability to cover the target workspace in the vertical direction. Spraying test steps were as follows: 1) The initial spraying height value was 100 mm; 2) Divide the seedling disk plane into m×n grids, and each grid point represented the spray position point, m=10, n=20; 3) The target path point group consisted of all the spray points at the current height, and the manipulator was controlled to pass through the target path point group sequentially; 4) The spraying height value was increased by 20 mm; 5) Repeat step 2), 3) and 4) until the spraying height value was equal to 1 020 mm. The spraying test mainly verified the manipulator’s ability to cover the target workspace in the horizontal direction. The high-speed video camera system was used to mark trajectory coordinates of manipulator in the motion planning test of carrying and spraying (high-speed camera was KODAK’s color CCD (charge coupled devices) camera, a resolution of 512×480 pixels, frame rate of 125 frames/s). Test results showed that the optimized manipulator could reach all limiting positions and other characteristic positions of target workspace, and the maximum relative positioning error was 0.98% which was within error range and could meet the accuracy requirements for manipulator containing the target workspace effectively; what was more, it was proved that the optimal parameters of manipulator were reasonable. Parameters optimization and experiment of three-dimesional seedling tray management robot could provide the reference for trajectory planning and motion control. [ABSTRACT FROM AUTHOR]