An effective and more efficient path planning algorithm is developed for a kinematically non-redundant free-floating space robot(FFSR) system by proposing a concept of degree of controllability(DOC) for underactuated systems. The DOC concept is proposed for making full use of the internal couplings and then achieving a better control effect, followed by a certain definition of controllability measurement which measures the DOC, based on obtaining an explicit and finite equivalent affine system and singular value decomposition. A simple method for nilpotent approximation of the Lie algebra generated by the FFSR system is put forward by direct Taylor expansion when obtaining the equivalent system. Afterwards, a large-controlla- bility-measurement(LCM) nominal path is searched by a weighted A* algorithm, and an optimal self-correcting method is designed to track the nominal path approximately, yielding an efficient underactuated path. The proposed strategy successfully avoids the drawback of inefficiency inherent in previous path-planning schemes, which is due to the neglect of internal couplings, and illustrative numerical examples show its efficacy. An effective and more efficient path planning algorithm is developed for a kinematically non-redundant free-floating space robot (FFSR) system by proposing a concept of degree of controllability (DOC) for underactuated systems. The DOC concept is proposed for making full use of the internal couplings and then achieve a better control effect, followed by a certain definition of controllability measurement which measures the measures the DOC, based on obtaining an explicit and finite equivalent affine system and singular value decomposition. A simple method for nilpotent approximation of the Lie algebra generated by the FFSR system is put forward by direct Taylor expansion when obtaining the equivalent system. Afterwards, a large-controlla- bility-measurement (LCM) nominal path is searched by a weighted A * algorithm, and an optimal self-correcting method is designed to track the nominal path approximately, yielding an efficient underactuated path. The proposed strategy successfully avoids the drawback of inefficiency inherent in previous path-planning schemes, which is due to the neglect of internal couplings, and illustrative numerical examples show its efficacy.