The current research on media convergence mostly divorces itself from the user's active subject vision, such as equating it with the superposition of media forms or the docking of technology. Media convergence should not only be regarded as content integration or user aggregation, but also as user-oriented intelligent scene matching. In terms of the practical dimension, media convergence needs to pay attention to "self-interested users" and enhance their interactivity and control. On the theoretical level, the user-oriented media convergence mechanism is represented by the new style of media convergence created by user needs and new media technologies, providing new services to users through intelligent scene matching and reshapes the dynamic evolution of media convergence in the process of sensory reorganization and new demand promotion. User-centered media convergence needs to reflect on recommendation "too appropriate" caused by personalized customization and other issues.