An optimal node competitive selection model is proposed to improve the accuracy and real-time of users’ optimal service selection. By analyzing hierarchical service model which is based on hierarchical data node, IP-domain based network users’ grouping model and time-sharing service evaluation based Region-users grouping model and definition are presented, a service evaluation based progressive grouping model is constructed, node competitive mechanism is introduced. On condition that load balancing is ensured, optimal node selection problem is solved, a selection model simulation is completed using MATLAB R2009a. Results show that, users can get optimal service through optimal node competitive selection model, compared to traditional models, users who can get optimal service grow linearly, the growth account for about 10% in all users. The efficiency of the model can satisfy users’ need.