Abstract: An inverse kinematics algorithm with improved tip following movement is proposed to solve the problems of expensive computational cost, exceeding joint limits and large configuration deviation for the inverse kinematics of hyper-redundant snake-arm robots. Constraints on the bending angles of snake arm are introduced to the tip-following method. And the update mode of the joint position is adjusted to make joints locate on the axis of the snake arm. By updating the spatial position of joints successively, the inverse kinematics of the hyper-redundant multi-section arm is converted into the inverse kinematics of a single-section arm with 2 degree-of-freedoms. Finally, the trajectory tracking effects of the snake-arm robot with the base moving, and also with the base fixed, are analyzed by simulation, and the performances of different methods under the same target position are compared. Results show that the improved algorithm can ensure that the bending angles do not exceed the given range and the amount of joint movement is diminishing from the tip to the base, yielding a more concordant movement. Compared with the Jacobian-based numerical method and the existing heuristic methods, the computational cost is decreased and the generated configuration deviation of the whole robot is reduced using the proposed method, which is more applicable for the real-time control of a snake-arm robot.