The oral microbiome plays vital roles in the human microbiome and human health. Although oral-microbiome dysbiosis can cause oral diseases and contribute to oral cancer, these relationships remain unclear. In this case-control study, we aimed to elucidate the link between the oral microbiome and potential mechanisms that can promote oral cancer progression. This study involved 1,022 participants, with the discovery and validation datasets including 637 patients (104 with oral cancer cases and 533 controls) and 385 patients (53 with oral cancer cases and 332 controls), respectively. Machine learning was employed, and the LightGBM model revealed four bacterial genera that were significantly associated with oral cancer. Streptococcus and Parvimonas were more abundant in the oral cancer group, whereas Corynebacterium and Prevotella showed decreased abundances. The levels of enzymes associated with fatty acid oxidation, such as carnitine O-palmitoyltransferase 1 (CPT1A), long-chain acyl-CoA synthetase, acyl-CoA dehydrogenase, diacylglycerol choline phosphotransferase, and H+-transporting ATPase, were significantly higher in patients with oral cancer than in control subjects. Streptococcus and Parvimonas correlated positively with the cytokines interleukin-6, tumor necrosis factor-alpha, and CPT1A, whereas Corynebacterium and Prevotella showed negative correlations. Our findings suggest a potential association between the changes in four distinct microbiomes and alterations in specific cytokines and enzymes in the context of oral cancer carcinogenesis. These microbiomes may function as promising predictive markers for oral cancer and improve its diagnostic accuracy.