In Recent days, raise of industrial loads and residential loads increasing rapidly. The raise of loads leads to power quality disturbances at the distribution side. These power quality disturbances cause blackout in the regular power supply, sometimes this may lead to hazardous fire accidents and loss of life and property. This is the main motivation to detect the power quality disturbances quickly and accurately. In this modern era, electrical engineers facing problem to detect and classify these PQ disturbances. This lead to many researchers work on this problem. In this paper, power quality disturbances are formulated and simulated. These PQ disturbances signals are analyzed through Discrete Wavelet Transform and classified through various supervised machine learning algorithms. Noise are introduced in to the PQD's and tested. All the simulations and programming carried out in MATLAB and Orange Data Mining Software.