Background <p>The bidirectional association between metabolic-associated fatty liver disease (MAFLD) and periodontitis (PD) is well-established, with periodontal pathogenic bacteria playing a key role. However, the differences and associations of microbial communities between subgingival plaque (core lesion) and saliva (non-invasive sample) in MAFLD patients with PD (MAFLD–PD) remain unclear. This study aims to characterize the microbial community profiles of these two microbial niches, evaluate the potential of saliva for non-invasive monitoring of pathogenic bacteria in subgingival plaque, and analyze their potential involvement in MAFLD–PD pathogenesis via functional pathway analysis.</p> Methods <p>A preliminary cross-sectional study included 10 MAFLD–PD patients, collected subgingival plaque and saliva, compared microbiota differences via 16S rRNA V3–V4 sequencing, analyzed inter-sample correlations of 4 key pathogenic bacteria and saliva’s predictive value, and predicted functional phenotypes using PICRUSt2 and BugBase.</p> Results <p>MAFLD–PD subgingival plaque samples (MP-SPG) and MAFLD–PD saliva samples (MP-SG) showed significant differences in alpha diversity (Shannon index, <i>p</i> = 0.02493) and beta diversity (PERMANOVA, <i>p</i> = 0.001). There were differences in community composition at both phylum and genus levels. Wilcoxon signed-rank test and linear discriminant analysis indicated that MP-SPG was predicted to be significantly enriched in core periodontal pathogens (such as <i>Porphyromonas gingivalis</i>), while MP-SG was significantly enriched in opportunistic pathogens (such as <i>Prevotella melaninogenica</i>). The total absolute abundance of the four pathogenic bacteria presented a strong positive correlation between saliva and subgingival plaque (Spearman <i>ρ</i> = 1.000). Given the very small sample size of this pilot study, this perfect correlation should be interpreted with extreme caution and regarded as a preliminary observation only, and no generalizable conclusion can be drawn at this stage. The total absolute abundance of these four pathogens in saliva exhibited preliminary predictive value (limited by small sample size) for their corresponding load in subgingival plaque (<i>R</i>^<i>2</i> = 0.8712). Functional prediction showed that MP-SPG was significantly enriched in the insulin resistance pathway (<i>p</i> = 0.00323), whereas MP-SG was significantly enriched in the lipopolysaccharide biosynthesis pathway (<i>p</i> = 0.02328). Phenotypic prediction indicated that MP-SPG was significantly enriched in the “Forms Biofilms” trait (<i>p</i> = 0.0491), while MP-SG was significantly enriched in the “Contains Mobile Elements” (<i>p</i> = 0.01973) and “Facultatively Anaerobic” (<i>p</i> = 0.03084) traits.</p> Conclusions <p>Significant niche differentiation was observed between the subgingival plaque and salivary microbiota of MAFLD–PD patients. This pilot study quantitatively characterized the differential characteristics of subgingival plaque and salivary microbiota in patients with MAFLD–PD, generated the preliminary hypothesis that the total absolute abundance of the four core pathogenic bacteria in saliva may serve as a candidate non-invasive predictive indicator for the pathogenic bacterial load in subgingival plaque, and explored, through in silico functional prediction analysis, the putative metabolic pathway characteristics of micro-organisms in different oral niches and their hypothesized potential roles in MAFLD–PD. These findings require validation in larger cohorts with appropriate control groups before clinical application.</p>