With the rapid development of China’s economy, the proportion of people’s disposable income is increasing. The consumption of daily necessities is increasing day by day. The traditional marketing channels of the existing clothing industry have been difficult to meet the development needs of clothing enterprises and consumer consumption needs. In response to this problem, this research introduces in-depth learning into the digital transformation of garment enterprises, focusing on how to combine in-depth learning with the innovation of marketing management mode in the digital transformation. This paper proposes an omni channel marketing management mode based on the improved PGBN user profile construction model and the improved RBF sales volume prediction model. Simulation experiments and comparative experiments are designed to test the effect of the sales volume prediction model and the new marketing management model. The test results show that the performance of the sales volume prediction model is improved by 15.97-19.87%. The new marketing management model improves the sales volume by more than 30%. The improvement effect is significant, which can provide an effective reference for the digital transformation of garment enterprises.