In the field of architecture, circulation refers to the way people move through and interact with a building. From the consumers' aspects, they may pay more attention on the relative position of different commodities and their demands. These behaviors can be deemed as the task-oriented works, which are less discussed comparing to the path planning algorithm. The task-oriented circulation planning may involve the dynamic allocation, demands, and resources, which may be more complex than the exists path problems. In this study, a task-oriented planning of circulation planning is proposed, which is modified and based on the Minimal Spanning Tree algorithm. Properties include circulation, turning points and exit points are involved in planning the optimal solutions based on different kinds of tasks. We also implemented and demonstrated a circulation planning on department of northern Taiwan, and try to provide some suggested circulations by using consumers' shopping lists as input. The proposed application can also be used to optimize the deployment of different commodities.