Substation disconnector are prone to severe dust adhesion and drying up of lubricating fluid after long-term operation, aiming at the problems of poor anti-interference ability and low accuracy of existing disconnector positioning and identification methods in complex scenarios, an accurate identification method of substation disconnector based on improved random forest algorithm is proposed. On the basis of the accurate positioning of disconnectors by combining decision tree algorithm and generalized Hough transform, the working status of disconnectors are classified by combining random forest algorithm and particle swarm optimization algorithm, so as to prepare for live cleaning of disconnectors in substations. Through the experimental comparison and analysis of the anti-jamming ability of the method and the recognition effect before and after the improvement, the superiority of the proposed method is verified. The experimental results show that this method can train the recognition model accurately, and the recognition effect is ideal, compared with the previous improvement, the recognition accuracy is improved by 9%, reaching 99.5%. Under the white noise environment, this method can still accurately identify the working state of three-phase disconnector, and the recognition accuracy is 97.5%.