Provided is an anti-electricity-theft early warning analysis method with single anomaly analysis. The method includes steps: screening and filtering invalid events in events generated by acquisition terminals and electric energy meters in a power distribution network: event screening of the acquisition terminals and the electric energy meters: screening and filtering the invalid events in the events generated by the acquisition terminals and the electric energy meters, and removing the invalid events; and acquisition data screening, and acquisition data screening of the acquisition terminals and the electric energy meters: single anomaly analysis. The effects of the method are that comprehensive judgment and analysis of anomaly information are conducted by employing various rules, near-real-time processing of massive data is realized with the combination of a big data mining technology, and on-line monitoring of on-site metering abnormal conditions and electricity-theft behaviors is performed; besides, according to various abnormal events and electricity-theft behaviors and an association relation among the kinds of the electricity-theft behaviors in a power consumption information acquisition system, comprehensive judgment and analysis are conducted with the combination of weights of various anomaly analysis models, the association relations among different models, and an association analysis algorithm, and the possibility of electricity theft for users is determined.