Aiming at the problem of low accuracy of current line selection technology in distribution network protection, A method of line selection protection based on power angle clustering analysis is proposed in this paper. A variety of fault characteristics of each feeder under different operation conditions are collected, and the historical database is formed. Through fuzzy clustering analysis, the clustering centers of fault and nonfault classes are obtained. The realtime data of each feeder is collected to form the characteristic samples to be tested. The power angle similarity is used to analyze the similarity between the characteristic samples to be tested and the historical characteristic samples, and the samples to be tested are classified into fault or non fault categories, so as to realize the accurate selection of fault feeders in distribution network. The effectiveness of the proposed protection method is verified by PSCAD. The simulation results shows that the misjudgment of single fault feature caused by external interference can be effectively eliminated by the proposed method. No need to set the setting value during the line selection process, which can accurately realize the fault line selection of the small current grounding system.