In the Muon package under study, the tracks are extrapolated using an algorithm which accounts for the magnetic field and the ionization (dE/dx). We improved the calculation of the field dependent term to increase the muon detection efficiency at lower momenta using a Runge-Kutta method. The muon identification and hadron separation in b-bbar jets is reported with the improved software. In the same framework, the utilization of the Kalman filter is introduced. The principle of the Kalman filter is described in some detail with the propagation matrix, with the Runge-Kutta term included, and the effect on low momenta single muons particles is described.