A new approach for global path planning of redundant manipulators is proposed. It poses the path planning problem as a finite time nonlinear control problem. The solution is found by a Newton-Raphson type algorithm. This technique is capable of handling various goal task descriptions as well as incorporating both joint and task space constraints. The algorithm has shown promising preliminary results in planning joint path sequences for 3R and 4R planar robots to meet Cartesian tip tracking and goal endpoint planning. It is robust with respect to local path planning problems such as singularity considerations and local minimum problems. Repetitive joint path solutions for cyclic end-effector tasks are also generated. Eventual goals of this work include implementation on full spatial robots, as well as provision of an interface for supervisory input to aid in path planning for more complex problems.