Investigating the mechanisms underlying microbial diversity is one of the challenges in microbiology. The dimension of diversity typically includes three aspects: taxonomic diversity (TD), phylogenetic diversity (PD), and functional diversity (FD). Anthropogenic activities, particularly those affecting groundwater ecosystems through contamination, represent an underexplored area of study. Microorganisms are crucial in mediating the effects of contaminants within these ecosystems, yet our understanding of their community responses remains partial. Thus, it is crucial to characterize the microbial community compositions, elucidate the relationship between biodiversity and ecosystem services, and explore the dynamics of microbial interactions with environmental pollutants for potential bioremediation applications. The Oak Ridge Field Research Center (OR-FRC) is one of the Department of Energy’s contamination sites with a variety of nutrients, stressors, and contaminants including uranium, nitrates, along with various volatile organic compounds. Detailed monitoring of hydrological and geological profiles of the OR-FRC site has made it an ideal location for investigating the reciprocal interactions between environmental conditions and microbial ecology and function. Microbial taxonomic diversity declined as the stress increased. However, whether the phylogenetic and functional diversities would show the same trend as the taxonomic diversity along the stress gradient remains unclear. We selected several groundwater wells with extremely high levels of nitrate, uranium, and extremely low pH from the OR-FRC site to answer these questions. Both taxonomic and phylogenetic α-diversities were declined in the most contaminated wells. In contrast, the decrease in functional α-diversity was modest and statistically insignificant, showing a better buffering capacity to environmental stress. Differences in functional composition, sometimes called β-diversity, were enlarged under high contaminated wells, while convergent functional composition was observed in uncontaminated wells. Relative abundances of most carbon degradation genes were decreased in contaminated wells, but those of many genes associated with nitrogen cycling, sulfur cycling, and metal homeostasis were increased. Environmental variables had a much higher explanatory power in functional composition than taxonomic and phylogenetic compositions, suggesting that niche selection favored microbial functionality. Together, we demonstrate that microbial functionality is more tolerant to stress than taxonomy and extend the Anna Karenina Principle based on Leo Tolstoy’s assertion in that microbial community adapts to a stressful environment in its own way. Since the functional composition is a sensitive and informative metric for evaluating the responses of microbial communities to environmental stress, and to further test the relationships between the functional genes’ complexity and ecosystems stability, we collected more groundwater samples at the OR-FRC site with more fluctuations in nitrate, pH and uranium. We used GeoChip data, a high-throughput functional gene array, to construct the functional molecular ecological networks. Notably, stress conditions led to decreased network complexity and stability, while network modularity increased. Functional genes associated with nitrogen cycling and metal homeostasis were significantly reduced under high contamination levels. We also identified deterministic assembly processes as key drivers of microbial community structure, although this trend was not obvious with escalating stress. Sulfate-reducing bacteria played an important role in the biogeochemistry cycles at the OR-FRC site, and potentially involved in the bioremediation of heavy metals and radionuclides. Thus, understanding their adaptation to the fluctuating environments are important. We used Desulfovibrio vulgaris Hildenborough, a model sulfate-reducing bacterium, to examine the effects of prior adaptation on evolutionary responses to elevated temperatures. Two groups of DvH populations with 5000 generations of experimental evolution under non-stress or salt stress conditions, and one group of DvH ancestral populations without previous experimental evolution history were evolved for 1000 generations under elevated temperature conditions. We found that most evolved populations showed increased growth rate and all evolved populations had increased fitness compared to their corresponding ancestor populations under heat stress conditions. Whole-genome sequencing indicated that significant difference of mutated genes was observed among three groups. These findings underscore the significance of evolutionary history in shaping microbial adaptation to new environmental challenges, with phenotypic convergence observed despite genetic divergence. Overall, this dissertation demonstrates that the linkage between microbial taxonomic and functional diversities is weakened in a polluted aquifer, environmental stresses decreased the complexity and stability of the functional molecular ecological networks. It advances our understanding of microbial ecology in contaminated environments and the adaptive mechanisms of microorganisms under varying stress conditions.