A method, apparatus and system are provided for controlling a scene structure of a mobile terminal for receiving and displaying data streams in a mobile broadcast system. When logical channels are multiplexed into one physical channel, the multiplexed channels are sequentially decoded and are displayed on a media. A channel circulation scene alternately displays the channels on one media in designated order. In a combination scene structure, a new node is added to a conventional Lightweight Application Scene Representation (LASeR) technology. To compose a scene, fields for time information and scene transition effect are provided. A delay time due to a channel change is reduced. A multi-channel media is effectively provided to address a problem when a user retrieves service content while directly switching a channel. The terminal dynamically composes the multi-channel media using stream source node information from a service provider without an additional mosaic channel.