This paper presents an improved designed induction motor, used for photovoltaic (PV) array fed water pumping system by modifying the rotor and stator slots. The overall system is designed without a mechanical sensor to reduce both cost and complexity with the simultaneous assurance of optimum power utilization of a PV array. The proposed system consists of an induction motor operated water pump, controlled by field-oriented control with artificial neural network current control technique. The maximum power point tracking as well as dc-link voltage, is regulated by three-phase voltage source inverter. The estimation of motor speed eliminates the use of a mechanical sensor and makes the system cheaper and robust. A new robust speed adaptive algorithm is presented, which is less dependent on parameters. A detailed study of various factors affecting the efficiency of the motor is given to improve the behavior of the induction motor drive for water pumping. The designed motor is tested on a developed prototype in the laboratory and its suitability is judged through various tests under steady-state and dynamic conditions of insolation variations.