In the process of three-dimensional (3D) path planning for robots, the efficiency of path planning is greatly affected by the algorithm itself. For the purpose of the shortest distance, a robot 3D path planning method was proposed based on ant colony parameter transfer algorithm. Ant colony algorithm was used to find the shortest path in the robot environment model which was established by grid method. For the parameter selection problem of the ant colony algorithm, the parameter transfer algorithm was used to obtain the optimal parameters. The known environment model and its corresponding ant colony optimal parameters were used as source tasks, and the source tasks were mapped to high-dimensional spaces. Connecting different source tasks through transfer parameters, the parameter transfer graph was established based on the knowledge of graph theory. The parameter transfer map was extended to include the target task, and a set of ant colony optimal parameters was assigned to the random unknown environment model. Simulation results showed that the ant colony algorithm based on parameter transfer can complete the robot 3D path planning quickly and effectively. Compared with the traditional parameter selection method and other intelligent optimization methods, the ant colony parameter transfer algorithm can greatly reduce the time required for path planning and improve the path planning performance.