Simultaneous Localization and Mapping is the one of essential techniques for mobile robot navigation. In this paper, we propose a fast SLAM method that uses particle weight based occupancy grid map, and also works well with in indoor environment. In addition, in the prediction step, our method that uses the polar scan matching and wheel odometry information improves the accuracy of prediction pose. In order to evaluate our proposed method, we use a simulation and the 2-wheeled mobile robot that is able to move in large-scale indoor environment. The experimental results show that our method is suitable for fast and accurate mapping and localization algorithm for mobile robot.