Background: We often experience visit-to-visit estimated glomerular filtration rate (eGFR) variability among patients with advanced chronic kidney disease (CKD). However, the impact of eGFR variability on renal prognosis is not well understood. Methods: The Kanagawa Valsartan Trial was a multicenter, randomized, open-label trial that compared the effect of add-on treatment with angiotensin II receptor blocker, valsartan, with that of conventional treatment on the rate of CKD progression in 293 advanced CKD patients. In this post hoc analysis, we compared the effect of high eGFR variability on end-stage renal disease (ESRD). To evaluate eGFR variability, we chose participants with ≥5 serial measurements of eGFR during the study period and assessed the residual coefficient of variation (CV) of eGFR (residual eGFR-CV) derived from a regression line of eGFR (eGFR slope). The primary end point was the first event of ESRD. Results: Among 237 patients with ≥5 serial measurements of eGFR in a median follow-up of 1.47 years, there were 54 ESRD events in the higher than median residual eGFR-CV group and 36 ESRD events in the lower than median eGFR-CV group (log-rank test, p = 0.008). In multivariate Cox regression analysis, high eGFR variability was significantly associated with the primary end point as well as hypertension, high proteinuria, low baseline eGFR and steep eGFR slope (hazards ratio 2.11, 95% CI 1.33-3.33, p = 0.001). Conclusion: High eGFR variability predicts poor renal prognosis; therefore, further attention is warranted for ambulatory patients with large eGFR fluctuations, especially those with advanced CKD.