In order to extract useful information from a large number of redundant and unreliable RFID (radio frequency identification) streaming data and improve the data quality of the RFID system, a filtering and cleaning method based on finite state machine for RFID streaming data is proposed. According to the experiment, suppressing interference tag data, filtering redundant tag data and extracting effective tag data can be realized in this method with reduced risk of system omission and false positives. Finally, using GIS technologies the monitoring and tracking results are displayed on the world map.