Abstract: In dead reckoning system, the limitation of complexity and instability of precision are often encountered for nonlinear problem solved using extended Kalman filtering (EKF). A effective method is provided using fully classical Kalman filtering formula for avoiding linearization process from data pre-processing. The data is processed before filter. Each measuring observation is transformed from polar coordinates into Cartesianism coordinates, then the filtering is done according to standard Kalman filtering in order to obtain best estimation. The validity of the method is tested by computer simulation.