Serum troponin T levels using a highly sensitive assay (hsTnT) in patients with chronic kidney disease (CKD) not on dialysis have not been examined. The aim of this prospective cohort study was to investigate the association of hsTnT with cardiac events in ambulatory CKD patients not on dialysis. The serum hsTnT level was measured in 442 ambulatory CKD patients not on dialysis whose estimated glomerular filtration rate was <60 ml/min/1.73 m². Patients were divided into quartiles according to hsTnT levels, and were followed up for 3 years. Cardiac events were defined as a cardiac death, acute myocardial infarction, unstable angina pectoris that required emergency coronary revascularization, or hospitalization for worsening heart failure. During the follow-up period (median 22 months), 63 cardiac events occurred. Kaplan–Meier incidence rates of cardiac events for 3 years were 0.88 %, 11.5 %, 19.0 %, and 41.4 % among quartiles of hsTnT levels (P < 0.0001). After adjusting for other confounders, elevated hsTnT level was an independent predictor for cardiac events (hazard ratio 6.18, 95 % confidence interval 1.38–27.7, P = 0.0080 for highest quartile vs lowest quartile). In addition, C-index for receiver-operating characteristic curves for cardiac events was greater in an established risks plus hsTnT model than in the established risk alone model (0.857 vs 0.844, P = 0.026). Using a highly sensitive assay, serum hsTnT level was shown to be an independent predictor of cardiac events and a promising risk stratification tool in patients with CKD not on dialysis.