This study presents bi input-extended Kalman filter (BI-EKF) based speed-sensorless direct torque control (DTC) of induction motors (IMs). For this aim, all states required for the speed-sensorless control system as well as commonly known parameter uncertainties related to stator resistance R s, rotor resistance R′r, and load torque t L including also viscous friction term have been estimated by using BI-EKF for a wide speed range. BI-EKF uses a single EKF algorithm with consecutive operation of two inputs obtained from two extended IM models developed for the simultaneous estimation of R′r and R s; thus, it has an advantage over previous EKF based studies utilizing two separate EKF algorithms for the same purpose. By assuming that stator phase voltages and currents are available, the proposed control system has been tested with the instantaneous step and/or linear variations of the velocity reference, tL, Rs, and R′r in simulations. In spite of those challenging variations, the control system performs quite well. ©2010 IEEE.