An indoor navigation system based on the RTOF (round time of flight) of Chirp signal is designed and implemented, which is used to locate and track multiple targets in a complex building interior. Time division multiplexing method to allow more tag positioning simultaneously, SDSTWR (symmetrical doublesided twoway ranging) algorithm is realized between mobile tag and anchor node, then tag sends the result to the positioning base station, and through to the computing center. Computing center uses the improved kalman filtering algorithm to reduce the NLOS error and the weighted correction three points localization algorithm to calculate the mobile tag location information, which is stored in the location server finally. The client is able to access to the server to check the result of appointed tag. The experiments show that, within certain complex floor structures, the precision can reach 2 m, meet the demand if indoor navigation accuracy. The system has also realized refresh time optimization with tag number, with flexible and deployable.