The UHF-RFID-based moving-target localization problem is investigated in this paper.In addition,a novel localization method,named B-AFK method,which integratesthe adaptive Kalman filter and the conventional BVIRE method,is presented.First,the UHF-RFID system is used to initialize the location of a mobile robot.Then,considering the uncertainties in the model and the statistical properties of the UHF-RFID localization system,an adaptive Kalman filter method is used to predict and update the robot's position.Further,an adaptive factor is introduced to compensate for the uncertainties in the noise covariance.Finally,a UHF-RFID-based localization platform is established and experiments are carried out to show that the proposed method outperforms the conventional BVIRE and linear Kalman filter in terms of precision and robustness.