In multiple target tracking, one of the crucial techniques is how to estimate unknown initiation points of tracks by incorporating a set of measurements under noisy and cluttered environment. Due to unknown data association and false alarms, all possible track initiation points need to be evaluated, which suffers from computational complexity. In this paper, we propose a computationally efficient track initiation method for multi-static multi-frequency passive coherent localization systems, where bistatic measurements from different illuminators are incorporated at a receiver to find the most probable track initiation points. By the proposed likelihood utility function, our method resolves unknown data association between measurements and targets to reduce computational complexity. In addition, the proposed method can be extended to process a set of multi-scan measurements in association-free manner, which results in improved estimation accuracy. Simulation results are provided to validate the performance of our proposed method.