This paper mainly studies the problem of autonomous exploration of mobile robots under indoor unknown environment, and proposes an improved strategy. Firstly, a method is proposed to extract the target point based on the accessible area to supplement the defects of the original method in the special environment structure. Then, a real-time topology map construction and optimization method based on laser data and grid map information is proposed, which makes the exploration map more streamlined and the exploration process is more efficient. Finally, the improved obstacle avoidance module is used to achieve the robot's motion control and reach the goal safetly. At the same time, the system adopts the distributed structure under robot operating system(ROS), allocates the whole algorithm to the client and the server reasonably, and reduces the hardware requirement of the system. Field experiments show that the proposed method has good performance, and offers advantages with respect to map building and obstacle avoidance in complex and unknown indoor environment.