The use of tools is a distinctive characteristic of human intelligence. However, traditional mapping algorithms from human hand to robot hand are ineffective in tool grasping due to the specific functional areas of tools. In this paper, we propose a hybrid mapping algorithm based on task division to address this problem. Firstly, we establish a human hand kinematic model and experimentally observe the synergistic effects of hand postures during the grasping of 44 common tools. Based on the grasping gestures, tools are categorized into multiple types, each exhibiting similar synergistic effects. Subsequently, human fingers are divided into grasping fingers and functional fingers according to task requirements. Grasping fingers are further classified into important and non-important fingers based on their significance in grasping tasks. Important fingers undergo joint configuration using object-based mapping, while the joints of other fingers are determined through posture synergies. Functional fingers require precise contact with tool regions, and this process is achieved through direct Cartesian mapping. Finally, we conduct comparative experiments to evaluate our algorithm, and the results demonstrate that our mapping algorithm effectively adapts to tool grasping tasks.