Diabetic kidney disease is a major cause of renal failure that urgently necessitates a breakthrough in disease management. Specific remedies are needed for preventing Type 2 diabetes which causes significant changes in an array of plasma metabolites. By untargeted metabolome analysis, phenyl sulfate (PS) increased with the progression of diabetes. In experimental diabetes models, PS administration induces albuminuria and podocyte damage due to the mitochondrial dysfunction. By clinical diabetic kidney disease (DKD) cohort analysis, it was also confirmed that the PS levels significantly correlate with basal and predicted 2-year progression of albuminuria. Phenol is synthesized from dietary tyrosine by gut bacterial-specific tyrosine phenol-lyase (TPL), and absorbed phenol is metabolized into PS in the liver. Inhibition of TPL reduces not only the circulating PS level but also albuminuria in diabetic mice. TPL inhibitor did not significantly alter the major composition, showing the non-lethal inhibition of microbial-specific enzymes has a therapeutic advantage, with lower selective pressure for the development of drug resistance. Clinically, 362 patients in a multi-center clinical study in diabetic nephropathy cohort (U-CARE) were analyzed with full data. The basal plasma PS level significantly correlated with ACR, eGFR, age, duration, HbA1c and uric acid, but not with suPAR. Multiple regression analysis revealed that ACR was the only factor that significantly correlated with PS. By stratified logistic regression analysis, in the microalbuminuria group, PS was the only factor related to the amount of change in the 2-year ACR in all models. PS is not only an early diagnosis marker, but also a modifiable cause and therefore a target for the treatment of DKD. Reduction of microbiota-derived phenol by the inhibitor should represent another aspect for developing drugs of DKD prevention.