Abstract: Mangrove wetlands are critical ‘blue carbon’ ecosystems that harbor diverse soil microbial communities playing essential roles in biogeochemical cycling. Within these communities, microorganisms can be categorized into generalists (broad ecological niches) and specialists (narrow ecological niches). However, the differential responses of these bacterial groups to infrastructure development, specifically bridge construction, and their underlying assembly mechanisms remain largely underexplored. This knowledge gap hinders our understanding of soil bacterial community evolution under anthropogenic disturbances and the development of effective mitigation strategies. Based on high-throughput 16S rRNA gene sequencing and bioinformatics analysis, this study investigated the impacts of bridge construction by comparing soil bacterial communities in two construction disturbance zones Sheet Pile Cofferdam (SP) and Steel Casing Pipe (SC) with those in an adjacent Undisturbed (UD) mangrove habitat. The results demonstrated a distinct divergence in how generalists and specialists responded to disturbance. In the undisturbed ecosystem, bacterial specialists exhibited significantly higher species richness (Mean=98.5) compared to generalists (Mean=49.4). However, bridge construction imposed a significant negative impact on specialists. Compared to the UD habitat, the species richness of specialists decreased notably by 38.73% in SP and 29.14% in SC sites, whereas generalist richness remained relatively stable. Moreover, the UD habitat harbored a higher number of indicator species than the SC and SP habitats, with 62.2% of them being specialists. Furthermore, construction activities profoundly altered the community structure of specialists (p<0.05), with soil organic carbon (SOC) and pH identified as the primary environmental drivers shaping these shifts. Community assembly analysis, utilizing the Modified Stochasticity Ratio (MST), revealed contrasting ecological strategies: generalist assembly was dominated by stochastic processes, suggesting high adaptability to environmental fluctuations; conversely, specialist assembly was governed predominantly by deterministic processes (environmental filtering). Notably, although bridge construction altered community composition and diversity, it did not change the community assembly processes of these bacterial groups. This indicates that the environmental stress induced by construction (e.g., carbon loss and pH alteration) acts as a strong filter, selectively eliminating taxa with narrow niche breadths. In conclusion, the impact of bridge construction on mangrove soil bacterial specialists was significantly greater than on generalists, highlighting the vulnerability of specialists and their potential utility as bio-indicators for monitoring anthropogenic disturbance. These findings provide a scientific basis for ecological management and biodiversity conservation. Therefore, conservation efforts during infrastructure development should prioritize minimizing soil physicochemical alterations to preserve the functional stability of specialist communities. Additionally, eco-friendly construction techniques, such as Steel Casing Pipes, which showed a relatively lower ecological impact compared to Sheet Pile Cofferdams, should be promoted to reduce environmental selection pressure on soil microorganisms.