Understanding microbial communities in mangrove ecosystems presents unique methodological challenges due to complex physicochemical gradients and varying biomass levels between water and sediment compartments. We systematically evaluated 16S rRNA amplicon and shotgun metagenomic sequencing approaches using the Dongzhai Port mangrove ecosystem as a model system. Metagenomic sequencing demonstrated superior taxonomic resolution, identifying 82 phyla compared to 52 phyla by 16S rRNA sequencing, and revealed distinct habitat-specific patterns in community structure. Both methods consistently showed higher bacterial diversity in sediment samples, reflecting enriched nutrient conditions and a stable physicochemical environment. While 16S rRNA sequencing effectively captured core community structures, metagenomic analysis provided deeper insights into functional capabilities, particularly in energy metabolism and nitrogen cycling pathways. Notably, metagenomic sequencing uncovered unique distribution patterns of key nitrogen metabolism genes (gltB, glnA/GLUL, gltD) between water and sediment compartments. Technical considerations, including PCR inhibition in sediments and variable biomass in water samples, significantly influenced method performance. Our findings provide a quantitative framework for method selection in mangrove microbiome studies, recommending 16S rRNA sequencing for broad community surveys and metagenomic sequencing for detailed functional analysis or rare taxa detection.