Power quality is a main subject to considerable attentions from utilities and customers owing to the popular uses of many non-linear electronic equipment in recent years. Harmonics, voltage swell, voltage sag, and, power interruption could downgrade the service quality. To ensure the power quality, detecting harmonic and voltage disturbances becomes an important issue. In other words, a detection method with classification capability will be helpful for detecting disturbances. The thesis proposed two models of power quality detection for power system disturbances using voltage-current(V-I) characteristics in the time domain with hybrid wavelets grey relational analysis (WGRA), and self-organizing feature map network (WSOM). Morlet wavelets are responsible for extracting features from voltages and currents. GRA and SOM were employed to identify the types of various disturbance patterns. Computer simulations have demonstrated the computational efficiency and accurate recognition capability for power quality detection and discrimination with an IEEE 14-Bus power system.