Tracking of maneuvering targets represents one of the basic casks to be carried out by radar and weapon systems. Usually the radar system is connected with a computer. The main problem involved in this case is the proper description of the target maneuver random process. Different models are used to represent the target maneuvering process. Simplest models should be used so as to reduce the burden on computer and to decrease the computation time to the minimum value permitted by the weapon system involved. Here is assumed a radar system with range and bearing measurements. The target is assumed to have constant velocity plus a random component that accounts for maneuvers. Different targets of different maneuverability are assumed, Augmented Kalman filter is assumed. The steady state behaviour of the system is obtaine,-1 through stu-ling the error covariance matrix in an off-line analysis. Convergence is also analyzed. Different radars detecting individual targets are considered and the performance of each radar is analysed to find the best performance conditions required such as the sampling time T.