A disturbances recognition method of power quality based on wavelet packet-energy entropy(WP-EE) was put forward,in which four layers wavelet packet decomposition of simulaed disturbance voltage signals were performed and characteristic vectors of wavelet packet energy entropy were extracted,principal components analysis(PCA) theory was used to extract characteristic vectors of wavelet packet of the voltage signal and the characteristic vectors were put into probabilistic neural network(PNN) for disturbance recognition.The method realizes optimum compression of disturbance data,simplifies structure of neural network classifier in disturbance classify,and enhances speed and precision of disturbance recognition.The simulation results showed the method has a very good disturbance recognition ability.