Abstract: It is well known that the positioning accuracy in cellular wireless communication systems is largely affected by the non-line-of-sight (NLOS) error，and thus the NLOS error mitigation algorithm has become a current hot subject of research in the wireless location field. This paper presents a robust location tracking algorithm for NLOS environments using the extended Kalman filtering (EKF) method. The algorithm firstly deduces the formula to compute the NLOS error under the least squares rule, and uses the constrained weighted least squares method directly to estimate the NLOS bias contained in each measurement range. A modified EKF tracking method based on NLOS correction is then applied to achieve a high accuracy for mobile location tracking. This method does not depend on a particular distribution of the NLOS error and need not the line-of-sight (LOS) or NLOS identification. Numerical results illustrate that the performance of the proposed algorithm is better than that of the classic EKF method and the iterated NLOS EKF algorithm. Even when base stations are all in NLOS condition, the proposed method can also meet the FCC requirement. Moreover, complexity experiments suggest that it supports most real-time applications.