Induction motor is one of the most widely used machines in industrial applications like water pumping system due to its high reliability, relatively low cost, and less maintenance requirements. Indirect Vector control of ac machine behaves similar to separately excited dc machine in which the torque and flux are controlled independently. Nevertheless, conventional vector-controlled induction motor drive has the disadvantage of required speed sensor and parameter sensitive, unknown variation during operation causes incorrect decoupling of flux and torque which leads to deterioration of drive performance. Sensor less speed control of induction machine has improvement of reducing the system size, cost high system reliability, long life span system and wide range of operating speed. The aim of this thesis is to design robust flux and speed estimation of induction motors based on NN by reducing the system components and achieve high dynamic performance. A neural network is an information processing system that is highly interconnected processing element working in unison used as an approach, for training the input error and change error and for estimator rotor flux and stator current used as an input of NN and rotor speed as an output. The performance of the proposed system has investigated through simulations by allowing for system response (at no-load and load condition) and speed tracking. The error of simulation result between actual and estimated speed have been approximately less than 0.20849% for transient response and 0.1122% speed tracking. the simulation result is established in MATLAB/SIMULINK. Keywords: Induction motor drives; indirect vector Control; PI controller; Feed forward neural network.