In this paper, an effective target detection and localization method is proposed for a passive bistatic radar (PBR) system. The PBR system consists of a commercial FM radio station, which is a non-cooperative illuminator of opportunity (IO), referred to as the transmitter antenna and multiple surveillance antennas that form an antenna array, e.g., uniform linear array (ULA). Unlike other literatures where the reference signal is received by a directional antenna, here, the reference signal (direct path) is estimated by beamforming method. Then a modified extensive cancellation algorithm (MECA) based on (least squares) LS method is proposed to solve the disturbance cancellation. After cancelling the disturbance, the matched filter (MF) and LS methods are used for range-Doppler estimation of targets, and then the angles of targets are estimated based on beamforming method. The proposed method is suitable for an antenna array. Simulation results are presented to illustrate the superiority of the proposed MECA disturbance cancellation method and parameter estimation method.