Automatic Dependent Surveillance-Broadcast (ADS-B) faces the safety challenge that is vulnerable to spoofing due to its open surveillance framework. In this paper, some ADS-B stations are set to position the aircraft by multilateration by use of Time Difference of Arrival (TDOA) and Time Sum of Arrival (TSOA) of ADS-B signals, and the false aircraft is eliminated by matching the calculated position of aircraft with the aircraft position decoding from ADS-B signals. Firstly, the time synchronization is achieved between ADS-B stations, the location model and algorithm is established based TDOA/TSOA. Then a modified Taylor algorithm using the Friedlander algorithm to get the initial value of the Taylor algorithm is proposed to calculate the plane position. The Monte Carlo simulation of four stations and five stations is carried out, the results show that the more accurate position can be achieved compared with the existing algorithms. And the five stations can achieve more accurate position compared with the four stations.