A multimedia document authoring system should provide analysis and validation tools that help authors find and correct mistakes before document deployment. Although very useful, multimedia validation tools are not often provided. Spatial validation of multimedia documents may be performed over the initial position of media items before presentation starts. However, such an approach does not lead to ideal results when media item placement changes over time. Some document authoring languages allow the definition of spatio-temporal relationships among media items and they can be moved or resized during runtime. Current validation approaches do not verify dynamic spatio-temporal relationships. This paper presents a novel approach for spatio-temporal validation of multimedia documents. We model the document state, extending the Simple Hyperme-dia Model (SHM), comprising media item positioning during the whole document presentation. Mapping between document states represent time lapse or user interaction. We also define a set of atomic formulas upon which the author's expectations related to the spatio-temporal layout can be described and analyzed.