At present, the "night economy" has become a highlight activity, and all businesses are working to meet the diversified market needs and improve the user's experience. While, the results are not satisfactory, and the total income is hard to increase. In this paper, multiple linear regression and support vector machines are used to study customer behavior, identify key influencing factors and predict customer behavior. The empirical research shows that, with the help of data mining technology, it can help enterprises to timely adjust advertising and marketing strategies to provide targeted service.