Renal tissue hypoxia is a final pathway in the development and progression of chronic kidney disease (CKD), but whether renal oxygenation predicts renal function decline in humans has not been proven. Therefore, we performed a prospective study and measured renal tissue oxygenation by blood oxygenation level-dependent magnetic resonance imaging (BOLD-MRI) in 112 patients with CKD, 47 with hypertension without CKD, and 24 healthy control individuals. Images were analyzed with the twelve-layer concentric objects method that divided the renal parenchyma in 12 layers of equal thickness and reports the mean R2* value of each layer (a high R2* corresponds to low oxygenation), along with the change in R2* between layers called the R2* slope. Serum creatinine values were collected to calculate the yearly change in estimated glomerular function rate (MDRD eGFR). Follow up was three years. The change in eGFR in CKD, hypertensive and control individuals was -2.0, 0.5 and -0.2 ml/min/1.73m/year, respectively. In multivariable regression analysis adjusted for age, sex, diabetes, RAS-blockers, eGFR, and proteinuria the yearly eGFR change correlated negatively with baseline 24 hour proteinuria and the mean R2* value of the cortical layers, and positively with the R2* slope, but not with the other covariates. Patients with CKD and high outer R2* or a flat R2* slope were three times more likely to develop an adverse renal outcome (renal replacement therapy or over a 30% increase in serum creatinine). Thus, low cortical oxygenation is an independent predictor of renal function decline. This finding should stimulate studies exploring the therapeutic impact of improving renal oxygenation on renal disease progression.