Abstract: Mostly, the personalized services meet the satisfaction of users with the same services under the same context. However, although these services have the same context may be different according to needs of users. Depending on environment around them each user is different from other in terms of their preferences. In this study, a user profile framework was enriched with user context and a set of ranked item features that selected with respect to how significant and how much information could extract from those features. The user profile features were chosen based on maximum weight sum model. In addition, the correlation among the selected features was calculated to leverage the relationships of the user profile’s constituents. As a result, the proposed user profile gives weighted features can be exploited for prediction purposes in future for several applications for instance; recommendation system.