The invention discloses a power distribution network state estimation abnormality recognition method and device and relates to the field of power distribution automation. In the prior art, manual problem searching is low in efficiency and long in period and has the defects that problem recognition is incomplete and the like, and the power distribution network measurement percent of pass cannot be effectively improved. The method comprises the steps that a power grid model and a cross section are automatically read from an SCADA system at regular time, a real-time network topology is generated, data are processed by the adoption of the power distribution network state estimation algorithm, disqualified measurement is recognized, equipment causing measurement abnormality is recognized according to preset rules, the owner of the equipment is judged, and a short message is sent to the owner, and operating workers are informed in time to carry out coherent processing. According to the technical scheme, manual checking and examining links are omitted, the cost is reduced, and efficiency is improved.