As IoT has been progressed, a lot of revolutions have been brought in various fields. Nowadays, Bluetooth Low Energy (BLE) beacons and smart devices are not so specific ones, but we can get considerable benefits by devising them. One of them is indoor location tracking. Methods using GPS is general concerning outdoor, however it is difficult for indoor to get GPS. Hence, we have focused upon BLE beacons, WIFI, and so on. Among them, BLE beacon is one of promising devices. It is because, BLE beacons are inexpensive, consume low energy, and are easy to install them in indoor location, such as supermarket. In this paper, we propose a method to take log data of historical RSSI and estimate indoor location tracking by mining data. In computational experiments, we take the log including accurate location labels automatically by using an automatic traveling robot, and make prediction model based on decision tree model using the data.