A 3-D single observer passive positioning method based on frequency variation, azimuth and elevation information is presented. According to the theory of particle kinematics, the radiation position of target can be obtained and processed by using EPT. The rough positioning result is corrected and smoothed to estimate the position of target. EPF (Extended Particle Filter) is a particle filtering method taking EKF as an important function. Simulation results show that EPF-based single observer passive position method is used to achieve more quick and stabilized convergences and better accuracy than conventional EKF. It has profound significance for improvement in passive positioning and tracking algorithm and actual application.