In so many circumstances, some sort of adaptation tuning flavors are highly recommended to improve the tracking filter performance. The efficacy of the most prevailing Kalman tracking filter based on Singer model has been presented in detail in this work. This popular tracking scheme is usually modeled by a third-order dynamic model in which the maneuver acceleration variable can be modeled by a first-order Gauss-Markov process through which the desired corresponding autocorrelation function and statistical characteristics can be attained. However, this tracking model incurred the problem of tracking both maneuvering and nonmaneuvering targets with the same degree of activity when the input excitation strength is chosen to be large. On the other hand, the filter performance becomes inaccurate and unacceptable and may loose the target when the same statistical attribute is chosen to be very small. Hencefore, a successful adaptive methodology is proposed to remedy such dilemmas and render the tracking filter performance to be sufficiently expedient against any sudden changes in the tracking scenario. This suggested technique facilitates the adaptation demands through investigating another appropriate dynamic model for the same problem at a hand, and by a justified addition of a bias term to account for the model discrepancy and ill matching. This unknown bias term is related to the target dynamic states by using an appropriately defined bias transition matrix, and hence can be processed and estimated through the use of another Kalman filter which entails a rather simply derived dynamic model for the bias term. This bias estimate is accordingly invested to correct the bias unknown state estimate through conducting recursive linear relations of the sensitivity matrices which are determined in a suitable fashion to reflect the effect of the bias estimate. The developed adaptive tracking technique, which will be called as " Bias-Removal (BR) technique", is evaluated using different target track conditions simulated on computer. The final results and comparisons, with Singer tracking model and bank of 3-Kalman adaptive approach, show great tracking accuracy improvements. Finally, the proposed adaptive tracking policy does not require a huge computational burdens, and thus seems to be a better candidate for the on-line critical applications than the aforementioned tracking techniques.