Abstract: Accurate classification identification of voltage dip disturbance is a prerequisite for power quality assessment and management.Most of the existing voltage dips’feature extraction consists of identifying and classifying single disturbance data.When using a mathematical transformation method for feature extraction,the data dimension is high and the amount of calculation is large.To solve these problems,a visual feature extraction and classification method based on a three-phase voltage space phasor model is proposed for multi-level voltage dips disturbance.First,the three-phase voltage waveform data are transformed into a spatial phasor model.Secondly,the voltage dip disturbances are clustered into visible circles or ellipses by using the K-mean algorithm.Finally,a logical regression algorithm is used to extract and classify the features of each cluster circle or ellipse.The simulation experiments for single disturbance and multi-level disturbance are done using the proposed method.The results show that the proposed method can effectively identify seven kinds of voltage-drop disturbances,such as A,C_a,C_b,C_c,D_a,D_b,D_c,etc.This method not only reduces the data dimension and the calculation amount of the model,but also reduces the risk of misidentification by eliminating the detection of the dynamic transition process.Overall it provides an effective means for the identification of multi-level voltage dip disturbances.