Induction machine, when driven by a field oriented controller behaves like a separately excited dc machine, where torque and flux are naturally decoupled and are controlled independently; and this control strategy allows high performance to be achieved from it.However, conventional vector controlled induction motor drive has the disadvantage of requiring speed sensor which increases the cost and complexity of the drive system.Moreover, drive performance is affected by rotor resistance, whose unknown variation during the operation causes incorrect decoupling of flux and torque which leads to deterioration of drive performance. This thesis focusses on the development of a high performance sensorless induction motor drive. Control of the drive is done in rotor flux coordinates which allows natural decoupling of flux and torque producing components of stator current space vector. Sensorless control of the drive system is achieved by developing rotor flux and speed estimation algorithms using only the measurable stator terminal quantities: the current and voltage. The state and parameter estimation problem is looked into as a question of proper mathematical modeling of the machine in an appropriate reference frame, which along with proper choice of states, enables us to develop an estimation algorithm capable of accurate estimation.In this work a two pronged approach is adopted to achieve the research objective. First approach is based on the fact that dynamics of rotor speed is much slower than that of electrical states of the machine. Hence, the time derivative of rotor speed can be conveniently equated to zero in the machine model used for implementing the estimation