Document-level relation extraction (DocRE) focuses on identifying semantic relations between entities that appear in different sentences. Existing methods, whether graph-based or sequence-based, predominantly rely on single-granularity representations, overlooking the fact that different relational triples often require distinct semantic granularities for accurate inference. Moreover, while bridge entities are frequently employed to facilitates the prediction, the contributions of non-bridge components remain largely underexplored in existing work. To address these limitations, we propose MNR (Multi-scale and Non-bridge Reasoning), a novel framework that introduces multi-scale semantic spaces and cross-axial attention mechanisms to enhance both relation extraction and relational reasoning. Experimental results on multiple widely adopted DocRE benchmarks verify both the effectiveness and generalization ability of our approach.