In view of the television program in 2012 fully digital, cable TV companies need to actively promote digital box for free, in order to increase the order rate of digital channels. Generally the cable TV companies did not analyze users’ preferences to launch marketing campaigns for digital channels. In this work, data mining techniques are used to analyze users’ preferences and discover implicit information to make channel recommendations for users. The RMF model is used to analyze users’ historical buying behavior. Then, K-means clustering algorithm is applied to cluster users who have similar buying behavior. For each user cluster, the channel recommendation is made according to the association and popularity of digital channels. The proposed channel recommendation method can help users quickly order their favorite channels, so that it brings a new business model for cable TV companies.