Continuum robots with floating bases demonstrate exceptional operational capabilities in confined spaces, such as those encountered in medical surgeries and equipment maintenance. However, developing low-cost solutions for their motion and planning problems remains a significant challenge in this field. This paper investigates the application of geometric iterative strategy methods to continuum robots, and proposes the algorithm based on an improved two-layer geometric iterative strategy for motion planning. First, we thoroughly study the kinematics and effective workspace of a multi-segment tendon-driven continuum robot with a floating base. Then, generalized iterative algorithms for solving arbitrary-segment continuum robots are proposed based on a series of problems such as initial arm shape dependence exhibited by similar methods when applied to continuum robots. Further, the task scenario is extended to a follow-the-leader task considering environmental factors, and further extended algorithm are proposed. Simulation comparison results with similar methods demonstrate the effectiveness of the proposed method in eliminating the initial arm shape dependence and improving the solution efficiency and accuracy. The experimental results further demonstrate that the method based on improved two-layer geometric iteration can be used for motion planning task of a continuum robot with a floating base, under an average deviation of about 4 mm in the end position, an average orientation deviation of no more than 1 degree, and the reduction of average number of iterations and time cost is 127.4 iterations and 72.6 ms compared with similar methods, respectively.