In the secondary side management ofa substation, the press-plate plays an important role. Aiming at the problem of unsatisfactory results of the existing accurate recognition methods for press-plate states, this study proposes a bilinear fine-grained press-plate state recognition method that incorporates the attention mechanism. First, the attention mechanism is used to focus attention on the contact part of the press-plate; then, the bilinear fine-grained model is used to focus on the key areas related to the opening and closing of the press-plate; finally, effective features are extracted from the key areas to achieve accurate recognition of the state of the press-plate. Experimental results show that the method can realize the end-to-end state recognition of the pressure plate, and the recognition accuracy can reach up to 98%. Compared to the traditional method, the accuracy and recall rate of the method are significantly improved.