Abstract: To address the problems of redundant search caused by random sampling and high computational cost of collision detection during path pruning in traditional sampling-based robotic manipulator path planning algorithms, a dual space–interval divide-and-conquer algorithmic framework was proposed to shorten path construction time and reduce pruning computation overhead, thereby improving operational efficiency. First, at the spatial divide-and-conquer level, an Octree-Connect algorithm was proposed by integrating an octree topological structure with a bidirectional search mechanism. A structured guided sampling strategy based on the expansion directions of octree child nodes was established to constrain high-probability feasible regions. Compared with traditional methods, the proposed approach reduced the number of sampled points by 44.3% and decreased initial planning time by 67.1%. Moreover, it maintained a high planning success rate even under extremely limited sampling budgets, demonstrating strong robustness. Second, at the interval divide-and-conquer level, a Binary Pruning algorithm was developed based on a recursive divide-and-conquer mechanism derived from binary search. By incorporating a greedy strategy that prioritizes connectivity checks between the start and goal nodes and intermediate nodes, the algorithm transformed global optimization with O(N) complexity into local convergence with O(log N) complexity, improving computational efficiency by 24.8% while ensuring path quality. Finally, through robotic manipulator motion simulations and multi-scenario validations in human-robot collaborative reducer assembly, the combined Octree-Connect + Binary Pruning approach was shown to stably reduce inverse kinematics solving time by 0.1~10 seconds and increase the planning success rate by at least 16.3%, thereby verifying the effectiveness and advanced nature of the proposed algorithmic combination.