Knowledge graphs (KGs), structured as multi-relational data of entities and relations, are essential for tasks like data analysis and recommendation systems. Knowledge graph completion (KGC), or link prediction, addresses KG incompleteness by inferring missing triples \((h, r, t)\) , and is critical for downstream applications. Region-based embedding models typically embed entities as points and relations as geometric regions to accomplish the task. Despite progress, these models often overlook the inherent semantic hierarchy in KGs. To address this limitation, we propose SectorE, a novel region-based embedding model utilizing polar coordinates. In SectorE, relations are modeled as annular sectors, leveraging both modulus and phase to capture inference patterns and relation attributes. Entities are embedded as points within these sectors, naturally encoding hierarchical structures. Experimental results on FB15k-237, WN18RR, and YAGO3-10 show that SectorE achieves competitive performance against various kinds of models, furthermore demonstrating strengths in semantic modeling capability.