Continuum robot is a new type of biomimetic robot,which realizes the motion by bending some parts of its body.So its path planning becomes more difficult even compared with hyper-redundant robots.In this paper a circular arc spline interpolating method is proposed for the robot shape description,and a new two-stage position-selectable-updating particle swarm optimization(TPPSO)algorithm is put forward to solve this path planning problem.The algorithm decomposes the standard PSO velocity’s single-step updating formula into twostage multi-point updating,specifically adopting three points as candidates and selecting the best one as the updated position in the first half stage,and similarly taking seven points as candidates and selecting the best one as the final position in the last half stage.This scheme refines and widens each particle’s searching trajectory,increases the updating speed of the individual best,and improves the converging speed and precision.Aiming at the optimization objective to minimize the sum of all the motion displacements of every segmental points and all the axial stretching or contracting displacements of every segment,the TPPSO algorithm is used to solve the path planning problem.The detailed solution procedure is presented.Numerical examples of five path planning cases show that the proposed algorithm is simple,robust,and efficient. Continuum robot is a new type of biomimetic robot, which realizes the motion by bending some parts of its body. So its path planning becomes more difficult even compared with hyper-redundant robots. This paper a circular arc spline interpolating method is proposed for the robot shape description, and a new two-stage position-selectable-updating particle swarm optimization (TPPSO) algorithm is put forward to solve this path planning problem. The algorithm decomposes the standard PSO velocity’s single-step updating formula into twostage multi-point updating , specifically using three points as candidates and selecting the best one as the updated position in the first half stage, and similar taking seven points as candidates and selecting the best one as the final position in the last half stage. This scheme refines and widens each particle’s searching trajectory, increases the updating speed of the individual best, and improves the converging speed and precision. Aiming at the optimization objective to minimize the sum of all the motion displacements of every segmental points and all the axial stretching or contracting displacements of every segment, the TPPSO algorithm is used to solve the path planning problem. The detailed solution procedure is presented. Numerical examples of five path planning cases show that the proposed algorithm is simple, robust, and efficient.