This paper describes a nine-dimensional extended Kalman filter (EKF) to jointly track and discriminate exoatmospheric active decoys in real time using motion features. By introducing an auxiliary variable as relative range delay ratio (RRDR), explicit nonlinear motion model of decoys in the East-North-Up coordinate system (ENU-CS) and the Jacobian matrix needed for the EKF are derived. The discrimination is then implemented according to the estimated deception range and its associated variance. The discrimination performance and the tracking accuracy compared to the traditional six-dimensional EKF are evaluated by Monte Carlo simulations. The advantage of the new algorithm lies in its real-time integration of tracking and discrimination.