This paper builds a viral marketing customer recommendation system in a location-based social network to find the most influential users. At present, social networks are growing rapidly, and messages are like a cold virus that can spread quickly through social networks. At present, the viral marketing customer recommendation system has low prediction accuracy, and we hope to enhance it to improve prediction accuracy. Currently, the viral marketing customer recommendation system only considers the customer's Network Value in the social network. The enhanced viral marketing customer recommendation system, in addition to considering the customer's Network Value, also increases considering the customer's interest, the customer's preference for the merchant and the customer's distance from the merchant, and can more accurately find the most influential users. Finally, the simulation proves that our method can improve the prediction accuracy and bring more visitors to the merchant.