The deep-sea ecosystem, a less-contaminated reservoir of antibiotic resistance genes (ARGs), has evolved antibiotic resistance for microbes to survive and utilize scarce resources. Research on the diversity and distribution of these genes in deep-sea environments is limited. Our metagenomics study employed short-read-based (SRB) and assembled-contig-based (ACB) methods to identify ARGs in deep-sea waters and sediments and assess their potential pathogenicity. SRB prediction was found to be more effective for studying the abundance and diversity of these genes, while combining both methods better illustrated the relationship of ARGs with the hosts. Deep-sea waters (DSW) and trenches had the highest diversity of ARGs, including β-lactams, multidrug resistance genes, and rifamycins. Mobile genetic elements, such as IncQ and RP4 plasmids, were also identified. The ratio of nonsynonymous to synonymous substitutions (pN/pS) values of these genes suggest different evolutionary strategies in response to deep-sea conditions and possible human impacts. These resistome profiles provide valuable insights into their natural origins as well as the ecological and evolutionary implications of antibiotic resistance in deep-sea ecosystems. The exploration of the global distribution of ARGs in diverse deep-sea environments is a novel approach that will assist in understanding their potential reservoirs and evolutionary mechanisms. Therefore, employing a comprehensive approach to studying ARGs is particularly necessary. Unique microbial life in deep-sea ecosystems, especially in deep-sea cold seeps sediments (DSCSS), deep-sea waters (DSW), and trench waters (TW), could be a valuable source of new antibiotics and resistance discovery.