In this study, Gordonia sp. HS126-4N is employed for dibenzothiophene (DBT) biodesulfurization, tracked over 9 days using SERS. During the initial lag phase, no significant spectral changes are observed, but after 48 h, elevated metabolic activity is evident. At 72 h, maximal bacterial population correlates with peak spectrum variance, followed by stable spectral patterns. Despite 2-hydroxybiphenyl (2-HBP) induced enzyme suppression, DBT biodesulfurization persists. PCA and PLS-DA analyze SERS spectra, revealing distinctive features linked to both bacteria and DBT, showcasing successful desulfurization and bacterial growth stimulation. PLS-DA achieves a specificity of 95.5 %, sensitivity of 94.3 %, and AUC of 74 %, indicating excellent classification of bacteria exposed to DBT. SERS effectively tracks DBT biodesulfurization and bacterial metabolic changes, offering insights into biodesulfurization mechanisms and bacterial development phases. This study highlights SERS' utility in biodesulfurization research, promising advancements in the field.