On-line highflux hemodiafiltration (HDF) is a clinically interesting and effective mode of renal replacement therapy, which offers the possibility to obtain an increased removal of both small and large solutes. The fundamental role of urea kinetic monitoring to assess dialysis adequacy in conventional hemodialysis has been widely studied. Both direct measurement of the urea removed by the modified direct dialysate quantitation (mDDQ) based on partial dialysate collection (PDC) and dialysate-based urea kinetic modeling (DUKM) using urea monitor have been advocated. The validity of this assessment tool in the patients with on-line HDF remained unclear. The aims of this investigation were (1) to compare the delivered Kt/V, urea mass removal (UMR), solute removal index (SRI) and normalized protein catabolic rate (nPCR) between pre- and postdilutional high-flux HDF; (2) to verify and compare the efficiency of pre- and postdilutional HDF using DUKM with on-line dialysate urea sensor, and mDDQ with partial dialysate collection. During both mode of HDF, the paired analysis urea removed and Kt/V showed no significant difference. Using mDDQ, mean values for predilutional mode were as follows: Kt/V 1.53 ± 0.01 UMR, 16.8 ± 0.3 g/session; urea clearance 178 ± 18 ml/min; SRI 75.5 ± 7.7%; urea distribution volume (V) 28.3 ± 1.2 liters; nPCR 1.34 ± 0.18 g/kg/day; on the other hand, mean values for postdilutional mode were Kt/V 1.58 ± 0.01; UMR 17.10 ± 0.28 g/session; urea clearance 184 ± 21 ml/min; SRI 77.2 ± 3.5%; urea distribution volume, 27.8 ± 1.5 liters; nPCR 1.34 ± 0.19 g/kg/day. The mean value of urea generation rate was 5.82 ± 1.12 mg/min during HDF. Our results showed that dialysis adequacy was achieved with both high-volume predilutional HDF and postdilutional HDF. These two modes of HDF provided similar and adequate small solute clearance. In addition, we found that on-line analysis of urea kinetics is a reliable tool for quantifying and assuring delivery of adequate dialysis.