This paper proposes a design and implementation of service robot for shopping in a supermarket, which accompanies customers and helps them to take the merchandises down from the shelf. And stop to answer the question from the customers during the patrolling in the mall. To create flexible interaction between customers and the service robot, graphic and verbal communications are investigated. Before the service scenario, the maps of the market are built through the simultaneous localization and mapping (SLAM) function with the position of merchandises obtained meanwhile. And the image features of merchandises are extracted by the recognition system. After receiving the request of grasping merchandises, service robot do path planning to optimize the moving distance during the task and grasp every object accurately. Further, to ensure the safety of customers, obstacle avoiding and emergency stop are investigated. With the proposed service robot system, a comfortable, efficient and labor saving market can be realized.