A novel approach has been made to calculate the switching angles for space vector modulation (SVM) technique, which is applied in a cycloinverter to generate high-frequency power output. The turn-on and turn-off times of SVM pattern are evaluated in such a way that the total harmonic distortion (THD) is minimised and the output voltage of cycloinverter approaches as close as possible to desired sinusoidal waveform. To accomplish this optimisation task, genetic algorithms (GAs) is used and THD is adopted in the proposed objective function. The lowest fitness value of this objective function decides the correct value of the switching angles. The switching strategy for the cycloinverter is presented on MATLABbased time-domain simulations. Simulation results show that the proposed approach with GAs, is superior to the conventional SVM. The output waveforms and their spectrum are presented to show effectiveness of the proposed scheme. Simulated results are also validated with the experimental results.