Abstract: As an important part of Internet finance, P2P lending has developed rapidly in recent years. But its potential risks are constantly exposed, which attracts attention in the industry. This paper researches an important problem in the supervision of P2P lending, which is the identification of systemically important platforms. According to the operation characteristics and operating environment of P2P lending platforms in China, an index system is designed. Afterwards, the corresponding weights are obtained by using BP neural network, so as to rank the systemic importance of the most representative P2P lending platforms in China. According to the ranking results, a batch of P2P lending platforms that need to be supervised and other relevant policy recommendations are put forward. The research has some theoretical significance and practical value for the supervision of P2P lending platforms.