This paper discusses an Artificial Neural Network (ANN) based method for reducing the torque ripple in a Doubly Fed Induction machine (DFIM) in which Direct Torque Control (DTC) scheme is applied. DFIM is gaining popularity as the best alternative for wind energy generation system. Often, wind energy generation demands good torque dynamic performance. ANN is used here to give pulses to the inverter in place of the conventional DTC scheme. The proposed method gives reduced torque ripple as compared to conventional DTC scheme.