Abstarct: In this thesis, Bearings-Only Tracking problem for moving surface target tracking is studied using a passive sensor. Main aspects of this practical problem are the un-observability of the target states, non-linearity of the measurement process, uncertainty of the target parameters, and correlation between the side angle measurement and the tracking dynamics for estimating the target distance. The innovations of this dissertation is to design an optimal observer for a nonlinear Bayesian problem, independent of the filters parameters and unknown parameters of the filter, with variety constraints of observer moving path via Chebyshev polynomials-baxsed problem modeling. Other achievment of this work is use of the interactive multiple model filters for solving the estimation problem and providing an appropriate indices for detection, as well as guaranteeing and matching of the online convergence of the estimator in case of visibility reduction. From research point of view related to the BOT problem, An optimal maneuvering of the observer is investigated and evaluated. In addition for introducing the performance measures of the nonlinear filter, the presented solutions using Monte Carlo simulation are provided and compared with the well-known relevant works.