Background/Aims: Frail patients with chronic kidney disease (CKD) have an increased hospitalisation and mortality rate. However, many popular frailty screening methods have not been validated in patients with CKD. This study evaluates the diagnostic accuracy of several frailty screening methods in patients with CKD G4–5 and those established on haemodialysis (G5D). Methods: Ninety participants with CKD G4–5D were recruited from Nephrology Outpatient Clinics and 2 Haemodialysis Units between December 2016 and December 2017. Frailty was diagnosed using the Fried Frailty Phenotype. The following frailty screening tests were evaluated: Clinical Frailty Scale, PRISMA-7, CKD Frailty Index, CKD FI-LAB, walking speed, hand grip strength and Short Physical Performance Battery. Results: The mean age of participants was 69 years (SD ±13). One-third of participants were dialysis-dependent. Nineteen (21%) patients were categorised as frail, 42 (47%) as pre-frail and 29 (32%) as robust. Overall, walking speed was the most discriminative measure (AUC 0.97 [95% CI 0.93–1.00], sensitivity 0.84 [95% CI 0.62–0.94], specificity 0.96 [95% CI 0.88–0.99]). The Clinical Frailty Scale had the best performance of the non-physical assessment frailty screening methods (AUC 0.90 [95% CI 0.84–0.97], sensitivity 0.79 [95% CI 0.57–0.91], specificity 0.87 [95% CI 0.78–0.93]). Conclusions: Walking speed can be used to accurately screen for frailty in CKD populations. If it is not practical to perform a physical assessment to screen for frailty, the Clinical Frailty Scale is a useful alternative.