This study aims to examine the divergence and convergence of cybersecurity legal terms in United States, China and European Union based on three specially created corpora of cybersecurity laws and regulations in those three jurisdictions. The corpus software Sketch Engine serves as a heuristic tool to construct the contextual meanings of legal terms in various jurisdictions. Through basic data analysis, it is found that information and data are core protected objects of cybersecurity laws and regulations and possess distinct meanings, which indicates they cannot be used interchangeably. Moreover, the types of information or data are regulated in a distinct manner in three corpora with the same information type interpreted in different ways, such as “copyright management information” and “sensitive information”. Therefore, the terminological interpretations of cybersecurity laws and regulations are spatially divergent with different interpretants existing in various jurisdictions, which calls for the reconstruction of cybersecurity legal terms when they are interpreted from one specific jurisdiction to another. This study also attempts to provide several pedagogical implications for how to employ the corpus-driven comparative approach to effectively facilitate the ESP teaching and learning concerning legal terms.