I. Introduction
With the standardized management of the telecom market and diversification of users, competitions between telecom operators become fierce increasingly. In order to seize market opportunities, attract more new customers and increase business revenue, telecom operators will introduce new service packages to seize market opportunities. However some service packages cost losses to the enterprise after a period of operation. Before the introduction of the new package, telecom enterprises can reduce operational risk and increase economic efficiency according to adjusting service packages timely through previewing new package accurately and estimating the number of users, cost, revenue, etc. This paper focuses on the method of dynamic preview of telecom services tariff, and researches prediction method of the amount of the new customers and the transfer customers with the use of theories and methods of Statistics and Data Mining. The research can help to preview telecom services tariff more accurately and reduce the risk of business operations, increase business revenue finally.