With the spread of COVID-19 in the whole world,the domestic economic development has been greatly affected, and the pressure of power supply company’s electricity recovery has been increasing. In order to solve the problem of poor accuracy of electricity charge recovery and less pertinence in the situation of COVID- 19, a risk analysis method based on long term and short term memory network is proposed. Firstly, the classification system of power supply users is established, and the classification of power supply users is realized by AP clustering. Secondly, the credit of power supply users is evaluated through the information of power supply users’credit investigation and judicial judgment. Thirdly,through the long-term and short-term memory network, combined with the user’s historical payment information and the user’s credit, the electricity charge recovery analysis is carried out, and the possible arrears risk is predicted. Finally, an example is run in a regional power supply company, and the results verify the feasibility and effectiveness of the proposed method.